data products

There are many different optimization techniques to choose from (see see sidebar, below ), but it is a well-understood field with robust and accessible solutions.

The Modeler takes the raw data and converts it into slightly more refined predicted data.

If we want a more sophisticated system, we can build another model for traffic congestion and yet another model to forecast weather conditions and their effect on the safest maximum speed.

By Jeremy Howard , Margit Zwemer and Mike Loukides

Sections

Download this free report

In the past few years, we’ve seen many data products based on predictive modeling.

Someone using Google’s self-driving car is completely unaware of the hundreds (if not thousands) of models and the petabytes of data that make it work.

Industrial engineers were among the first to begin using neural networks, applying them to problems like the optimal design of assembly lines and quality control.

For motor vehicle traffic, IBM performed a project with the city of Stockholm to optimize traffic flows that reduced congestion by nearly a quarter, and increased the air quality in the inner city by 25%.

Their actuaries could build models to predict a customer’s likelihood of being in an accident and the expected value of claims.

Step 4 of the Drivetrain Approach for Google is now part of tech history: Larry Page and Sergey Brin invented the graph traversal algorithm PageRank and built an engine on top of it that revolutionized search.

Back in 1997, AltaVista was king of the algorithmic search world.

We are entering the era of data as drivetrain, where we use data not just to generate more data (in the form of predictions), but use data to produce actionable outcomes.

There may be one detailed model for mechanical systems, a separate model for thermal systems, and yet another for electrical systems, etc.

We could just build a simple model of distance / speed-limit to predict arrival time with little more than a ruler and a road map.

Picture a Model Assembly Line for data products that transforms the raw data into an actionable outcome.

He went into Strand bookstore in New York City and asked for a book similar to Toni Morrison’s “Beloved.”

Any city with metered stoplights already has all the necessary information; they just haven’t found a way to suck the meaning out of it.

In another area where objective-based data products have the power to change lives, the CMU extension in Silicon Valley has an active project for building data products to help first responders after natural or man-made disasters .

There may be one detailed model for mechanical systems, a separate model for thermal systems, and yet another for electrical systems, etc.

The

data is in the wing materials’ physical properties; costs are listed in another tab of the application.

These products range from weather forecasting to recommendation engines to services that predict airline flight times more accurately than the airline itself.

Models developed to simulate fluid dynamics and turbulence have been applied to improving traffic and pedestrian flows by using the placement of exits and crowd control barriers as levers.

Zafu’s approach is not to send their customers directly to the clothes, but to begin by asking a series of simple questions about the customers’ body type, how well their other jeans fit, and their fashion preferences.

The vehicle needs to use a simulator to examine the results of the possible actions it could take.

For an insurance company, policy price is the product, so an optimal pricing model is to them what the assembly line is to automobile manufacturing.

These days, it is trivial to use some type of heuristic search algorithm to predict the drive times along various routes (a

Simulator ) and then pick the shortest one (an

Optimizer ) subject to constraints like avoiding bridge tolls or maximizing gas mileage.

By Jeremy Howard , Margit Zwemer and Mike Loukides

Sections

Download this free report

In the past few years, we’ve seen many data products based on predictive modeling.

Sidebar: Optimization in the real world

Optimization is a classic problem that has been studied by Newton and Gauss all the way up to mathematicians and engineers in the present day.

Making the wrong choices comes at a cost to the retailer in the form of reduced margins (discounts that do not drive extra sales), opportunity costs for the scarce real-estate on their homepage (taking up space in the recommendation feed with products the customer doesn’t like or would have bought without a recommendation) or the customer tuning out (sending so many unhelpful email promotions that the customer filters all future communications as spam).

They also considered inputs outside of their control, like competitors’ strategies, macroeconomic conditions, natural disasters, and customer “stickiness.”

Insurers have centuries of experience in prediction, but as recently as 10 years ago, the insurance companies often failed to make optimal business decisions about what price to charge each new customer.

The data collection and recommendation steps are not an add-on; they are Zafu’s entire business model — women’s jeans are now a data product.

(“If Hulu shows me that same dog food ad one more time, I’m gonna stop watching!”)

The takeaway, whether you are a tiny startup or a giant insurance company, is that we unconsciously use optimization whenever we decide how to get to where we want to go.

Engineers are often quietly on the leading edge of algorithmic applications because they have long been thinking about their own modeling challenges in an objective-based way.

This will require conducting many randomized experiments in order to collect data about a wide range of recommendations for a wide range of customers.

These days, it is trivial to use some type of heuristic search algorithm to predict the drive times along various routes (a

Simulator ) and then pick the shortest one (an

Optimizer ) subject to constraints like avoiding bridge tolls or maximizing gas mileage.

The danger in this hill-climbing approach is that if the steps are too small, we may get stuck at one of the many local maxima in the foothills, which will not tell us the best set of controllable inputs.

Making the wrong choices comes at a cost to the retailer in the form of reduced margins (discounts that do not drive extra sales), opportunity costs for the scarce real-estate on their homepage (taking up space in the recommendation feed with products the customer doesn’t like or would have bought without a recommendation) or the customer tuning out (sending so many unhelpful email promotions that the customer filters all future communications as spam).

He went into Strand bookstore in New York City and asked for a book similar to Toni Morrison’s “Beloved.”

Optimization is a process we are all familiar with in our daily lives, even if we have never used algorithms like gradient descent or simulated annealing.

The danger in this hill-climbing approach is that if the steps are too small, we may get stuck at one of the many local maxima in the foothills, which will not tell us the best set of controllable inputs.

For motor vehicle traffic, IBM performed a project with the city of Stockholm to optimize traffic flows that reduced congestion by nearly a quarter, and increased the air quality in the inner city by 25%.

It is easy to stumble into the trap of thinking that since data exists somewhere abstract, on a spreadsheet or in the cloud, that data products are just abstract algorithms.

We will show a systematic approach to step 4 that doesn’t require a PhD in computer science.

For motor vehicle traffic, IBM performed a project with the city of Stockholm to optimize traffic flows that reduced congestion by nearly a quarter, and increased the air quality in the inner city by 25%.

For an insurance company, policy price is the product, so an optimal pricing model is to them what the assembly line is to automobile manufacturing.

The girl behind the counter recommended William Faulkner’s “Absolom Absolom.”

To jump-start this process, we suggest a four-step approach that has already transformed the insurance industry.

A company like Amazon represents every purchase that has ever been made as a giant sparse matrix, with customers as the rows and products as the columns.

Industrial engineers were among the first to begin using neural networks, applying them to problems like the optimal design of assembly lines and quality control.

The technology exists to build data products that can revolutionize entire industries.

We will show a systematic approach to step 4 that doesn’t require a PhD in computer science.

These outcomes can be fed to an

Optimizer to build a functioning and cost-effective airplane wing.

Suppose we wanted to get from San Francisco to the Strata 2012 Conference in Santa Clara .

Because the simulation is at a per-policy level, the insurer can view the impact of a given set of price changes on revenue, market share, and other metrics over time.

We don’t claim that the Drivetrain Approach is the best or only method; our goal is to start a dialog within the data science and business communities to advance our collective vision.

We could just build a simple model of distance / speed-limit to predict arrival time with little more than a ruler and a road map.

The danger in this hill-climbing approach is that if the steps are too small, we may get stuck at one of the many local maxima in the foothills, which will not tell us the best set of controllable inputs.

The operator can adjust the input levers to answer specific questions like, “What will happen if our company offers the customer a low teaser price in year one but then raises the premiums in year two?”

For an insurance company, policy price is the product, so an optimal pricing model is to them what the assembly line is to automobile manufacturing.

Jeannie Stamberger of Carnegie Mellon University Silicon Valley explained to us many of the possible applications of predictive algorithms to disaster response, from text-mining and sentiment analysis of tweets to determine the extent of the damage, to swarms of autonomous robots for reconnaissance and rescue, to logistic optimization tools that help multiple jurisdictions coordinate their responses.

In another area where objective-based data products have the power to change lives, the CMU extension in Silicon Valley has an active project for building data products to help first responders after natural or man-made disasters .

The objective is to escape a recommendation filter bubble , a term which was originally coined by Eli Pariser to describe the tendency of personalized news feeds to only display articles that are blandly popular or further confirm the readers’ existing biases.

They also considered inputs outside of their control, like competitors’ strategies, macroeconomic conditions, natural disasters, and customer “stickiness.”

The profit for a very low price will be in the red by the value of expected claims in the first year, plus any overhead for acquiring and servicing the new customer.

Why not bundle simulation and optimization engines with a physical engine, all inside the black box of a car?

But those models did not solve the pricing problem, so the insurance companies would set a price based on a combination of guesswork and market studies.

Sidebar: Optimization in the real world

Optimization is a classic problem that has been studied by Newton and Gauss all the way up to mathematicians and engineers in the present day.

The levers are the vehicle controls we are all familiar with: steering wheel, accelerator, brakes, etc.

It was necessary to build this dataset by randomly changing the prices of hundreds of thousands of policies over many months.

We don’t claim that the Drivetrain Approach is the best or only method; our goal is to start a dialog within the data science and business communities to advance our collective vision.

He went into Strand bookstore in New York City and asked for a book similar to Toni Morrison’s “Beloved.”

Someone using Google’s self-driving car is completely unaware of the hundreds (if not thousands) of models and the petabytes of data that make it work.

Amazon’s recommendation engine is probably the best one out there, but it’s easy to get it to show its warts.

The price elasticity model is a curve of price versus the probability of the customer accepting the policy conditional on that price.

In another area where objective-based data products have the power to change lives, the CMU extension in Silicon Valley has an active project for building data products to help first responders after natural or man-made disasters .

Prediction technology can be interesting and mathematically elegant, but we need to take the next step.

Making the wrong choices comes at a cost to the retailer in the form of reduced margins (discounts that do not drive extra sales), opportunity costs for the scarce real-estate on their homepage (taking up space in the recommendation feed with products the customer doesn’t like or would have bought without a recommendation) or the customer tuning out (sending so many unhelpful email promotions that the customer filters all future communications as spam).

The objective is to escape a recommendation filter bubble , a term which was originally coined by Eli Pariser to describe the tendency of personalized news feeds to only display articles that are blandly popular or further confirm the readers’ existing biases.

Then, Google came along and transformed online search by beginning with a simple question: What is the user’s main objective in typing in a search query?

The best way to illustrate this process is with a familiar data product: search engines.

[ Note: Co-author Jeremy Howard founded ODG.

He went into Strand bookstore in New York City and asked for a book similar to Toni Morrison’s “Beloved.”

Jeremy Howard examined these questions in his Strata CA 12 session, “ From Predictive Modelling to Optimization: The Next Frontier .”

These products range from weather forecasting to recommendation engines to services that predict airline flight times more accurately than the airline itself.

Data scientists now have the predictive tools to build products that increase the common good, but they need to be aware that building the models is not enough if they do not also produce optimized, implementable outcomes.

These days, it is trivial to use some type of heuristic search algorithm to predict the drive times along various routes (a

Simulator ) and then pick the shortest one (an

Optimizer ) subject to constraints like avoiding bridge tolls or maximizing gas mileage.

If it makes a right turn at 55 mph in these weather conditions, will it skid off the road?

In another area where objective-based data products have the power to change lives, the CMU extension in Silicon Valley has an active project for building data products to help first responders after natural or man-made disasters .

The screenshot below is taken from a model integration tool designed by Phoenix Integration .

Models developed to simulate fluid dynamics and turbulence have been applied to improving traffic and pedestrian flows by using the placement of exits and crowd control barriers as levers.

Plenty of websites sell designer denim, but for many women, high-end jeans are the one item of clothing they never buy online because it’s hard to find the right pair without trying them on.

By Jeremy Howard , Margit Zwemer and Mike Loukides

Sections

Download this free report

In the past few years, we’ve seen many data products based on predictive modeling.

The levers are the vehicle controls we are all familiar with: steering wheel, accelerator, brakes, etc.

Great predictive modeling is an important part of the solution, but it no longer stands on its own; as products become more sophisticated, it disappears into the plumbing.

Instead of the femme-bot voice of the GPS unit telling us which route to take and where to turn, what would it take to build a car that would make those decisions by itself?

We can build a

Simulator to test the utility of each of the many possible books we have in stock, or perhaps just over all the outputs of a collaborative filtering model of similar customer purchases, and then build a simple

Optimizer that ranks and displays the recommended books based on their simulated utility.

Google realized that the objective was to show the most relevant search result; for other companies, it might be increasing profit, improving the customer experience, finding the best path for a robot, or balancing the load in a data center.

The Strand bookseller made a brilliant but far-fetched recommendation probably based more on the character of Morrison’s writing than superficial similarities between Morrison and other authors.

The Simulator’s result is fed to an

Optimizer , which takes the surface of possible outcomes and identifies the highest point.

The

objective is clearly defined: build an airplane wing.

Click to enlarge .

The self-driving car needs to take the next step: after

simulating all the possibilities, it must

optimize the results of the simulation to pick the best combination of acceleration and braking, steering and signaling, to get us safely to Santa Clara.

The

data is in the wing materials’ physical properties; costs are listed in another tab of the application.

Prediction technology can be interesting and mathematically elegant, but we need to take the next step.

For example, a pair of jeans that is often paired with a particular top, or the first part of a series of novels that often leads to a sale of the whole set.

Making the wrong choices comes at a cost to the retailer in the form of reduced margins (discounts that do not drive extra sales), opportunity costs for the scarce real-estate on their homepage (taking up space in the recommendation feed with products the customer doesn’t like or would have bought without a recommendation) or the customer tuning out (sending so many unhelpful email promotions that the customer filters all future communications as spam).

These models predicted whether customers would renew their policies in one year, allowing for changes in price and willingness to jump to a competitor.

In the future, we hope to see optimization taught in business schools as well as in statistics departments.

The girl behind the counter recommended William Faulkner’s “Absolom Absolom.”

Many optimization procedures are iterative; they can be thought of as taking a small step, checking our elevation and then taking another small uphill step until we reach a point from which there is no direction in which we can climb any higher.

This encompasses all the interactions that a retailer has with its customers outside of the actual buy-sell transaction, whether making a product recommendation, encouraging the customer to check out a new feature of the online store, or sending sales promotions.

These disaster applications are a particularly good example of why data products need simple, well-designed interfaces that produce concrete recommendations.

There are many techniques to avoid this problem, some based on statistics and spreading our bets widely, and others based on systems seen in nature, like biological evolution or the cooling of atoms in glass.

Here is a screenshot of the “Customers Who Bought This Item Also Bought” feed on Amazon from a search for the latest book in Terry Pratchett’s “ Discworld series :”

All of the recommendations are for other books in the same series, but it’s a good assumption that a customer who searched for “Terry Pratchett” is already aware of these books.

Then, Google came along and transformed online search by beginning with a simple question: What is the user’s main objective in typing in a search query?

The objective is to escape a recommendation filter bubble , a term which was originally coined by Eli Pariser to describe the tendency of personalized news feeds to only display articles that are blandly popular or further confirm the readers’ existing biases.

Once they have the data in this format, data scientists apply some form of collaborative filtering to “fill in the matrix.”

What is most important about these examples is that the engineers who designed these data products didn’t start by building a neato robot and then looking for something to do with it.

For example, if customer A buys products 1 and 10, and customer B buys products 1, 2, 4, and 10, the engine will recommend that A buy 2 and 4.

Here is a screenshot of the “Customers Who Bought This Item Also Bought” feed on Amazon from a search for the latest book in Terry Pratchett’s “ Discworld series :”

All of the recommendations are for other books in the same series, but it’s a good assumption that a customer who searched for “Terry Pratchett” is already aware of these books.

Here is a screenshot of the “Customers Who Bought This Item Also Bought” feed on Amazon from a search for the latest book in Terry Pratchett’s “ Discworld series :”

All of the recommendations are for other books in the same series, but it’s a good assumption that a customer who searched for “Terry Pratchett” is already aware of these books.

Jeannie Stamberger of Carnegie Mellon University Silicon Valley explained to us many of the possible applications of predictive algorithms to disaster response, from text-mining and sentiment analysis of tweets to determine the extent of the damage, to swarms of autonomous robots for reconnaissance and rescue, to logistic optimization tools that help multiple jurisdictions coordinate their responses.

The objective is to escape a recommendation filter bubble , a term which was originally coined by Eli Pariser to describe the tendency of personalized news feeds to only display articles that are blandly popular or further confirm the readers’ existing biases.

The

objective is clearly defined: build an airplane wing.

Plenty of websites sell designer denim, but for many women, high-end jeans are the one item of clothing they never buy online because it’s hard to find the right pair without trying them on.

The objective of a recommendation engine is to drive additional sales by surprising and delighting the customer with books he or she would not have purchased without the recommendation .

Then, Google came along and transformed online search by beginning with a simple question: What is the user’s main objective in typing in a search query?

There may be one detailed model for mechanical systems, a separate model for thermal systems, and yet another for electrical systems, etc.

The four steps in the Drivetrain Approach.

So, why aren’t we building them?

Engineers start by defining a clear

objective : They want a car to drive safely from point A to point B without human intervention.

As one engineer on the Google self-driving car project put it in a recent Wired article , “We’re analyzing and predicting the world 20 times a second.”

This has improved emergency evacuation procedures for subway stations and reduced the danger of crowd stampedes and trampling during sporting events.

She cut through the chaff of the obvious to make a recommendation that will send the customer home with a new book, and returning to Strand again and again in the future.

We can keep the “like” model that we have already built as well as the causality model for purchases with and without recommendations, and then take a staged approach to adding additional models that we think will improve the marketing effectiveness.

For motor vehicle traffic, IBM performed a project with the city of Stockholm to optimize traffic flows that reduced congestion by nearly a quarter, and increased the air quality in the inner city by 25%.

We call it the

Drivetrain Approach , inspired by the emerging field of self-driving vehicles.

There is a

Modeler for aerodynamics and mechanical structure that can then be fed to a

Simulator to produce the Key Wing Outputs of cost, weight, lift coefficient and induced drag.

Any city with metered stoplights already has all the necessary information; they just haven’t found a way to suck the meaning out of it.

Someone using Google’s self-driving car is completely unaware of the hundreds (if not thousands) of models and the petabytes of data that make it work.

Someone using Google’s self-driving car is completely unaware of the hundreds (if not thousands) of models and the petabytes of data that make it work.

For example, if customer A buys products 1 and 10, and customer B buys products 1, 2, 4, and 10, the engine will recommend that A buy 2 and 4.

Google realized that the objective was to show the most relevant search result; for other companies, it might be increasing profit, improving the customer experience, finding the best path for a robot, or balancing the load in a data center.

The models will take both the levers and any uncontrollable variables as their inputs; the outputs from the models can be combined to predict the final state for our objective.

We can keep the “like” model that we have already built as well as the causality model for purchases with and without recommendations, and then take a staged approach to adding additional models that we think will improve the marketing effectiveness.

Jeannie Stamberger of Carnegie Mellon University Silicon Valley explained to us many of the possible applications of predictive algorithms to disaster response, from text-mining and sentiment analysis of tweets to determine the extent of the damage, to swarms of autonomous robots for reconnaissance and rescue, to logistic optimization tools that help multiple jurisdictions coordinate their responses.

The price elasticity model is a curve of price versus the probability of the customer accepting the policy conditional on that price.

The objective of a recommendation engine is to drive additional sales by surprising and delighting the customer with books he or she would not have purchased without the recommendation .

By Jeremy Howard , Margit Zwemer and Mike Loukides

Sections

Download this free report

In the past few years, we’ve seen many data products based on predictive modeling.

Jeannie Stamberger of Carnegie Mellon University Silicon Valley explained to us many of the possible applications of predictive algorithms to disaster response, from text-mining and sentiment analysis of tweets to determine the extent of the damage, to swarms of autonomous robots for reconnaissance and rescue, to logistic optimization tools that help multiple jurisdictions coordinate their responses.

Brian Ripley’s seminal book on pattern recognition gives credit for many ideas and techniques to largely forgotten engineering papers from the 1970s.

Only after these first three steps do we begin thinking about building the predictive

models .

I think of this as a complicated machine (full-system) where the curtain is withdrawn and you get to model each significant part of the machine under controlled experiments and then simulate the interactions.

We call it the

Drivetrain Approach , inspired by the emerging field of self-driving vehicles.

For an insurance company, policy price is the product, so an optimal pricing model is to them what the assembly line is to automobile manufacturing.

To jump-start this process, we suggest a four-step approach that has already transformed the insurance industry.

Sidebar: Optimization in the real world

Optimization is a classic problem that has been studied by Newton and Gauss all the way up to mathematicians and engineers in the present day.

Prediction technology can be interesting and mathematically elegant, but we need to take the next step.

The

objective is clearly defined: build an airplane wing.

A great image for optimization in the real world comes up in a recent TechZing podcast with the co-founders of data-mining competition platform Kaggle .

Their actuaries could build models to predict a customer’s likelihood of being in an accident and the expected value of claims.

The first component of ODG’s Modeler was a model of price elasticity (the probability that a customer will accept a given price) for new policies and for renewals.

For example, resistance in the electrical system produces heat, which needs to be included as an input for the thermal diffusion and cooling model.

Making the wrong choices comes at a cost to the retailer in the form of reduced margins (discounts that do not drive extra sales), opportunity costs for the scarce real-estate on their homepage (taking up space in the recommendation feed with products the customer doesn’t like or would have bought without a recommendation) or the customer tuning out (sending so many unhelpful email promotions that the customer filters all future communications as spam).

The profit for a very low price will be in the red by the value of expected claims in the first year, plus any overhead for acquiring and servicing the new customer.

But these products are still just making predictions, rather than asking what action they want someone to take as a result of a prediction.

We will show how to go about building an optimized marketing strategy that mitigates these effects.

Suppose we wanted to get from San Francisco to the Strata 2012 Conference in Santa Clara .

These firms have plenty of experience building models of each of the components and systems in their final product, whether they’re building a server farm or a fighter jet.

We could construct a patience model for the customers’ tolerance for poorly targeted communications: When do they tune them out and filter our messages straight to spam?

ODG approached this problem with an early use of the Drivetrain Approach and a practical take on step 4 that can be applied to a wide range of problems.

The expected profit curve is just a slice of the surface of possible outcomes.

Then, Google came along and transformed online search by beginning with a simple question: What is the user’s main objective in typing in a search query?

The

data is in the wing materials’ physical properties; costs are listed in another tab of the application.

Many optimization procedures are iterative; they can be thought of as taking a small step, checking our elevation and then taking another small uphill step until we reach a point from which there is no direction in which we can climb any higher.

Making the wrong choices comes at a cost to the retailer in the form of reduced margins (discounts that do not drive extra sales), opportunity costs for the scarce real-estate on their homepage (taking up space in the recommendation feed with products the customer doesn’t like or would have bought without a recommendation) or the customer tuning out (sending so many unhelpful email promotions that the customer filters all future communications as spam).

What we would really like to do is emulate the experience of Mark Johnson, CEO of Zite , who gave a perfect example of what a customer’s recommendation experience should be like in a recent TOC talk .

This will require conducting many randomized experiments in order to collect data about a wide range of recommendations for a wide range of customers.

This has improved emergency evacuation procedures for subway stations and reduced the danger of crowd stampedes and trampling during sporting events.

We don’t claim that the Drivetrain Approach is the best or only method; our goal is to start a dialog within the data science and business communities to advance our collective vision.

To jump-start this process, we suggest a four-step approach that has already transformed the insurance industry.

Making the wrong choices comes at a cost to the retailer in the form of reduced margins (discounts that do not drive extra sales), opportunity costs for the scarce real-estate on their homepage (taking up space in the recommendation feed with products the customer doesn’t like or would have bought without a recommendation) or the customer tuning out (sending so many unhelpful email promotions that the customer filters all future communications as spam).

From there, they developed an optimized pricing process that added hundreds of millions of dollars to the insurers’ bottom lines.

ODG approached this problem with an early use of the Drivetrain Approach and a practical take on step 4 that can be applied to a wide range of problems.

The operator can adjust the input levers to answer specific questions like, “What will happen if our company offers the customer a low teaser price in year one but then raises the premiums in year two?”

Instead, let’s design an improved recommendation engine using the Drivetrain Approach, starting by reconsidering our

objective .

By Jeremy Howard , Margit Zwemer and Mike Loukides

Sections

Download this free report

In the past few years, we’ve seen many data products based on predictive modeling.

Google realized that the objective was to show the most relevant search result; for other companies, it might be increasing profit, improving the customer experience, finding the best path for a robot, or balancing the load in a data center.

Step 4 of the Drivetrain Approach for Google is now part of tech history: Larry Page and Sergey Brin invented the graph traversal algorithm PageRank and built an engine on top of it that revolutionized search.

The third step was to consider what new

data they would need to produce such a ranking; they realized that the implicit information regarding which pages linked to which other pages could be used for this purpose.

This situation changed in 1999 with a company called Optimal Decisions Group (ODG).

Instead of the femme-bot voice of the GPS unit telling us which route to take and where to turn, what would it take to build a car that would make those decisions by itself?

But these products are still just making predictions, rather than asking what action they want someone to take as a result of a prediction.

The third step was to consider what new

data they would need to produce such a ranking; they realized that the implicit information regarding which pages linked to which other pages could be used for this purpose.

ODG’s competitors use different techniques to find an optimal price, but they are shipping the same over-all data product.

The four steps in the Drivetrain Approach.

Making the wrong choices comes at a cost to the retailer in the form of reduced margins (discounts that do not drive extra sales), opportunity costs for the scarce real-estate on their homepage (taking up space in the recommendation feed with products the customer doesn’t like or would have bought without a recommendation) or the customer tuning out (sending so many unhelpful email promotions that the customer filters all future communications as spam).

Then, Google came along and transformed online search by beginning with a simple question: What is the user’s main objective in typing in a search query?

They can also explore how the distribution of profit is shaped by the inputs outside of the insurer’s control: “What if the economy crashes and the customer loses his job?

To jump-start this process, we suggest a four-step approach that has already transformed the insurance industry.

Multiplying these two curves creates a final curve that shows price versus expected profit (see Expected Profit figure, below).

Merely predicting what will happen isn’t good enough.

This encompasses all the interactions that a retailer has with its customers outside of the actual buy-sell transaction, whether making a product recommendation, encouraging the customer to check out a new feature of the online store, or sending sales promotions.

But these products are still just making predictions, rather than asking what action they want someone to take as a result of a prediction.

The takeaway, whether you are a tiny startup or a giant insurance company, is that we unconsciously use optimization whenever we decide how to get to where we want to go.

The operator can adjust the input levers to answer specific questions like, “What will happen if our company offers the customer a low teaser price in year one but then raises the premiums in year two?”

Google realized that the objective was to show the most relevant search result; for other companies, it might be increasing profit, improving the customer experience, finding the best path for a robot, or balancing the load in a data center.

Simple: we want to optimize the lifetime value from each customer.

Step 4 of the Drivetrain Approach for Google is now part of tech history: Larry Page and Sergey Brin invented the graph traversal algorithm PageRank and built an engine on top of it that revolutionized search.

This encompasses all the interactions that a retailer has with its customers outside of the actual buy-sell transaction, whether making a product recommendation, encouraging the customer to check out a new feature of the online store, or sending sales promotions.

Then, Google came along and transformed online search by beginning with a simple question: What is the user’s main objective in typing in a search query?

But these products are still just making predictions, rather than asking what action they want someone to take as a result of a prediction.

The takeaway, whether you are a tiny startup or a giant insurance company, is that we unconsciously use optimization whenever we decide how to get to where we want to go.

The danger in this hill-climbing approach is that if the steps are too small, we may get stuck at one of the many local maxima in the foothills, which will not tell us the best set of controllable inputs.

The second component of ODG’s Modeler related price to the insurance company’s profit, conditional on the customer accepting this price.

But as data scientists build increasingly sophisticated products, they need a systematic design approach.

Nest is designing smart thermostats that learn the home-owner’s temperature preferences and then optimizes their energy consumption.

There may be one detailed model for mechanical systems, a separate model for thermal systems, and yet another for electrical systems, etc.

As predictive modeling and optimization become more vital to a wide variety of activities, look out for the engineers to disrupt industries that wouldn’t immediately appear to be in the data business.

Instead, let’s design an improved recommendation engine using the Drivetrain Approach, starting by reconsidering our

objective .

The profit for a very low price will be in the red by the value of expected claims in the first year, plus any overhead for acquiring and servicing the new customer.

While the insurers were reluctant to conduct these experiments on real customers, as they’d certainly lose some customers as a result, they were swayed by the huge gains that optimized policy pricing might deliver.

These days, it is trivial to use some type of heuristic search algorithm to predict the drive times along various routes (a

Simulator ) and then pick the shortest one (an

Optimizer ) subject to constraints like avoiding bridge tolls or maximizing gas mileage.

]

ODG identified which

levers the insurance company could control: what price to charge each customer, what types of accidents to cover, how much to spend on marketing and customer service, and how to react to their competitors’ pricing decisions.

He went into Strand bookstore in New York City and asked for a book similar to Toni Morrison’s “Beloved.”

There is a

Modeler for aerodynamics and mechanical structure that can then be fed to a

Simulator to produce the Key Wing Outputs of cost, weight, lift coefficient and induced drag.

We don’t claim that the Drivetrain Approach is the best or only method; our goal is to start a dialog within the data science and business communities to advance our collective vision.

This curve moves from almost certain acceptance at very low prices to almost never at high prices.

But these products are still just making predictions, rather than asking what action they want someone to take as a result of a prediction.

Full video from that session is embedded below:

Related:

© 2012, O'Reilly Media, Inc.

(800) 889-8969 or (707) 827-7019 Monday-Friday 7:30am-5pm PT

All trademarks and registered trademarks appearing on oreilly.com are the property of their respective owners.

Finally, ODG started to design the

models that could be used to optimize the insurer’s profit.

I think of this as a complicated machine (full-system) where the curtain is withdrawn and you get to model each significant part of the machine under controlled experiments and then simulate the interactions.

I think of this as a complicated machine (full-system) where the curtain is withdrawn and you get to model each significant part of the machine under controlled experiments and then simulate the interactions.

She cut through the chaff of the obvious to make a recommendation that will send the customer home with a new book, and returning to Strand again and again in the future.

Sidebar: Optimization in the real world

Optimization is a classic problem that has been studied by Newton and Gauss all the way up to mathematicians and engineers in the present day.

There are many techniques to avoid this problem, some based on statistics and spreading our bets widely, and others based on systems seen in nature, like biological evolution or the cooling of atoms in glass.

We could just build a simple model of distance / speed-limit to predict arrival time with little more than a ruler and a road map.

But those models did not solve the pricing problem, so the insurance companies would set a price based on a combination of guesswork and market studies.

By Jeremy Howard , Margit Zwemer and Mike Loukides

Sections

Download this free report

In the past few years, we’ve seen many data products based on predictive modeling.

Models developed to simulate fluid dynamics and turbulence have been applied to improving traffic and pedestrian flows by using the placement of exits and crowd control barriers as levers.

Full video from that session is embedded below:

Related:

© 2012, O'Reilly Media, Inc.

(800) 889-8969 or (707) 827-7019 Monday-Friday 7:30am-5pm PT

All trademarks and registered trademarks appearing on oreilly.com are the property of their respective owners.

She cut through the chaff of the obvious to make a recommendation that will send the customer home with a new book, and returning to Strand again and again in the future.

While their models were good at finding relevant websites, the answer the user was most interested in was often buried on page 100 of the search results.

While their models were good at finding relevant websites, the answer the user was most interested in was often buried on page 100 of the search results.

While their models were good at finding relevant websites, the answer the user was most interested in was often buried on page 100 of the search results.

Data science is beginning to pervade even the most bricks-and-mortar elements of our lives.

ODG approached this problem with an early use of the Drivetrain Approach and a practical take on step 4 that can be applied to a wide range of problems.

There are many different optimization techniques to choose from (see see sidebar, below ), but it is a well-understood field with robust and accessible solutions.

By Jeremy Howard , Margit Zwemer and Mike Loukides

Sections

Download this free report

In the past few years, we’ve seen many data products based on predictive modeling.

They also considered inputs outside of their control, like competitors’ strategies, macroeconomic conditions, natural disasters, and customer “stickiness.”

It is easy to stumble into the trap of thinking that since data exists somewhere abstract, on a spreadsheet or in the cloud, that data products are just abstract algorithms.

We don’t claim that the Drivetrain Approach is the best or only method; our goal is to start a dialog within the data science and business communities to advance our collective vision.

But these products are still just making predictions, rather than asking what action they want someone to take as a result of a prediction.

Prediction technology can be interesting and mathematically elegant, but we need to take the next step.

Making the wrong choices comes at a cost to the retailer in the form of reduced margins (discounts that do not drive extra sales), opportunity costs for the scarce real-estate on their homepage (taking up space in the recommendation feed with products the customer doesn’t like or would have bought without a recommendation) or the customer tuning out (sending so many unhelpful email promotions that the customer filters all future communications as spam).

They also considered inputs outside of their control, like competitors’ strategies, macroeconomic conditions, natural disasters, and customer “stickiness.”

But these products are still just making predictions, rather than asking what action they want someone to take as a result of a prediction.

The wing box includes the design

levers like span, taper ratio and sweep.

Note here the different levels: models of individual components, tied together in a simulation given a set of inputs, iterated through over different input sets in a search optimizer.”

I think of this as a complicated machine (full-system) where the curtain is withdrawn and you get to model each significant part of the machine under controlled experiments and then simulate the interactions.

The objective of a recommendation engine is to drive additional sales by surprising and delighting the customer with books he or she would not have purchased without the recommendation .

On Amazon, the top results for a similar query leads to another book by Toni Morrison and several books by well-known female authors of color.

This encompasses all the interactions that a retailer has with its customers outside of the actual buy-sell transaction, whether making a product recommendation, encouraging the customer to check out a new feature of the online store, or sending sales promotions.

Insurers have centuries of experience in prediction, but as recently as 10 years ago, the insurance companies often failed to make optimal business decisions about what price to charge each new customer.

We don’t claim that the Drivetrain Approach is the best or only method; our goal is to start a dialog within the data science and business communities to advance our collective vision.

He went into Strand bookstore in New York City and asked for a book similar to Toni Morrison’s “Beloved.”

Google realized that the objective was to show the most relevant search result; for other companies, it might be increasing profit, improving the customer experience, finding the best path for a robot, or balancing the load in a data center.

We don’t know what design approaches will be developed in the future, but right now, there is a need for the data science community to coalesce around a shared vocabulary and product design process that can be used to educate others on how to derive value from their predictive models.

The third step was to consider what new

data they would need to produce such a ranking; they realized that the implicit information regarding which pages linked to which other pages could be used for this purpose.

There is a

Modeler for aerodynamics and mechanical structure that can then be fed to a

Simulator to produce the Key Wing Outputs of cost, weight, lift coefficient and induced drag.

But why not think bigger?

Irfan Ahmed of CloudPhysics provides a good taxonomy of predictive modeling that describes this entire assembly line process:

“When dealing with hundreds or thousands of individual components models to understand the behavior of the full-system, a ‘search’ has to be done.

What is most important about these examples is that the engineers who designed these data products didn’t start by building a neato robot and then looking for something to do with it.

Amazon’s recommendation engine is probably the best one out there, but it’s easy to get it to show its warts.

On Amazon, the top results for a similar query leads to another book by Toni Morrison and several books by well-known female authors of color.

It will be low in cases where the algorithm recommends a familiar book that the customer has already rejected (both components are small) or a book that he or she would have bought even without the recommendation (both components are large and cancel each other out).

There may be one detailed model for mechanical systems, a separate model for thermal systems, and yet another for electrical systems, etc.

He went into Strand bookstore in New York City and asked for a book similar to Toni Morrison’s “Beloved.”

The first component of ODG’s Modeler was a model of price elasticity (the probability that a customer will accept a given price) for new policies and for renewals.

The first component of ODG’s Modeler was a model of price elasticity (the probability that a customer will accept a given price) for new policies and for renewals.

Insurers have centuries of experience in prediction, but as recently as 10 years ago, the insurance companies often failed to make optimal business decisions about what price to charge each new customer.

They began by defining the

objective that the insurance company was trying to achieve: setting a price that maximizes the net-present value of the profit from a new customer over a multi-year time horizon, subject to certain constraints such as maintaining market share.

Many optimization procedures are iterative; they can be thought of as taking a small step, checking our elevation and then taking another small uphill step until we reach a point from which there is no direction in which we can climb any higher.

To jump-start this process, we suggest a four-step approach that has already transformed the insurance industry.

Then, Google came along and transformed online search by beginning with a simple question: What is the user’s main objective in typing in a search query?

Only after these first three steps do we begin thinking about building the predictive

models .

While their models were good at finding relevant websites, the answer the user was most interested in was often buried on page 100 of the search results.

Suppose we wanted to get from San Francisco to the Strata 2012 Conference in Santa Clara .

What choice are we actually helping him or her make?

By Jeremy Howard , Margit Zwemer and Mike Loukides

Sections

Download this free report

In the past few years, we’ve seen many data products based on predictive modeling.

They also considered inputs outside of their control, like competitors’ strategies, macroeconomic conditions, natural disasters, and customer “stickiness.”

Nest is designing smart thermostats that learn the home-owner’s temperature preferences and then optimizes their energy consumption.

While their models were good at finding relevant websites, the answer the user was most interested in was often buried on page 100 of the search results.

I think of this as a complicated machine (full-system) where the curtain is withdrawn and you get to model each significant part of the machine under controlled experiments and then simulate the interactions.

What is most important about these examples is that the engineers who designed these data products didn’t start by building a neato robot and then looking for something to do with it.

There are many techniques to avoid this problem, some based on statistics and spreading our bets widely, and others based on systems seen in nature, like biological evolution or the cooling of atoms in glass.

A great image for optimization in the real world comes up in a recent TechZing podcast with the co-founders of data-mining competition platform Kaggle .

From there, they developed an optimized pricing process that added hundreds of millions of dollars to the insurers’ bottom lines.

The danger in this hill-climbing approach is that if the steps are too small, we may get stuck at one of the many local maxima in the foothills, which will not tell us the best set of controllable inputs.

As one engineer on the Google self-driving car project put it in a recent Wired article , “We’re analyzing and predicting the world 20 times a second.”

The wing box includes the design

levers like span, taper ratio and sweep.

The levers are the vehicle controls we are all familiar with: steering wheel, accelerator, brakes, etc.

While the insurers were reluctant to conduct these experiments on real customers, as they’d certainly lose some customers as a result, they were swayed by the huge gains that optimized policy pricing might deliver.

Step 4 of the Drivetrain Approach for Google is now part of tech history: Larry Page and Sergey Brin invented the graph traversal algorithm PageRank and built an engine on top of it that revolutionized search.

The operator can adjust the input levers to answer specific questions like, “What will happen if our company offers the customer a low teaser price in year one but then raises the premiums in year two?”

It was necessary to build this dataset by randomly changing the prices of hundreds of thousands of policies over many months.

Plenty of websites sell designer denim, but for many women, high-end jeans are the one item of clothing they never buy online because it’s hard to find the right pair without trying them on.

What if a 100-year flood hits his home?

The objective is to escape a recommendation filter bubble , a term which was originally coined by Eli Pariser to describe the tendency of personalized news feeds to only display articles that are blandly popular or further confirm the readers’ existing biases.

A look inside the Modeler.

Zafu can tailor their recommendations to fit as well as their jeans because their system is asking the right questions.

Many optimization procedures are iterative; they can be thought of as taking a small step, checking our elevation and then taking another small uphill step until we reach a point from which there is no direction in which we can climb any higher.

The objective is to escape a recommendation filter bubble , a term which was originally coined by Eli Pariser to describe the tendency of personalized news feeds to only display articles that are blandly popular or further confirm the readers’ existing biases.

They began by defining the

objective that the insurance company was trying to achieve: setting a price that maximizes the net-present value of the profit from a new customer over a multi-year time horizon, subject to certain constraints such as maintaining market share.

We could just build a simple model of distance / speed-limit to predict arrival time with little more than a ruler and a road map.

Instead, let’s design an improved recommendation engine using the Drivetrain Approach, starting by reconsidering our

objective .

Making the wrong choices comes at a cost to the retailer in the form of reduced margins (discounts that do not drive extra sales), opportunity costs for the scarce real-estate on their homepage (taking up space in the recommendation feed with products the customer doesn’t like or would have bought without a recommendation) or the customer tuning out (sending so many unhelpful email promotions that the customer filters all future communications as spam).

There is a

Modeler for aerodynamics and mechanical structure that can then be fed to a

Simulator to produce the Key Wing Outputs of cost, weight, lift coefficient and induced drag.

Someone using Google’s self-driving car is completely unaware of the hundreds (if not thousands) of models and the petabytes of data that make it work.

He went into Strand bookstore in New York City and asked for a book similar to Toni Morrison’s “Beloved.”

Quite a few.

But these products are still just making predictions, rather than asking what action they want someone to take as a result of a prediction.

In general, when choosing an objective function to optimize, we need less emphasis on the “function” and more on the “objective.”

For motor vehicle traffic, IBM performed a project with the city of Stockholm to optimize traffic flows that reduced congestion by nearly a quarter, and increased the air quality in the inner city by 25%.

Step 4 of the Drivetrain Approach for Google is now part of tech history: Larry Page and Sergey Brin invented the graph traversal algorithm PageRank and built an engine on top of it that revolutionized search.

The next machine on the assembly line is a

Simulator , which lets ODG ask the “what if” questions to see how the levers affect the distribution of the final outcome.

Next, we consider what

data the car needs to collect; it needs sensors that gather data about the road as well as cameras that can detect road signs, red or green lights, and unexpected obstacles (including pedestrians).

Full video from that session is embedded below:

Related:

© 2012, O'Reilly Media, Inc.

(800) 889-8969 or (707) 827-7019 Monday-Friday 7:30am-5pm PT

All trademarks and registered trademarks appearing on oreilly.com are the property of their respective owners.

The operator can adjust the input levers to answer specific questions like, “What will happen if our company offers the customer a low teaser price in year one but then raises the premiums in year two?”

What we would really like to do is emulate the experience of Mark Johnson, CEO of Zite , who gave a perfect example of what a customer’s recommendation experience should be like in a recent TOC talk .

These products range from weather forecasting to recommendation engines to services that predict airline flight times more accurately than the airline itself.

Once we have these models, we construct a

Simulator and an

Optimizer and run them over the combined models to find out what recommendations will achieve our objectives: driving sales and improving the customer experience.

From there, they developed an optimized pricing process that added hundreds of millions of dollars to the insurers’ bottom lines.

We are entering the era of data as drivetrain, where we use data not just to generate more data (in the form of predictions), but use data to produce actionable outcomes.

We are entering the era of data as drivetrain, where we use data not just to generate more data (in the form of predictions), but use data to produce actionable outcomes.

By Jeremy Howard , Margit Zwemer and Mike Loukides

Sections

Download this free report

In the past few years, we’ve seen many data products based on predictive modeling.

The danger in this hill-climbing approach is that if the steps are too small, we may get stuck at one of the many local maxima in the foothills, which will not tell us the best set of controllable inputs.

Making the wrong choices comes at a cost to the retailer in the form of reduced margins (discounts that do not drive extra sales), opportunity costs for the scarce real-estate on their homepage (taking up space in the recommendation feed with products the customer doesn’t like or would have bought without a recommendation) or the customer tuning out (sending so many unhelpful email promotions that the customer filters all future communications as spam).

They began by defining the

objective that the insurance company was trying to achieve: setting a price that maximizes the net-present value of the profit from a new customer over a multi-year time horizon, subject to certain constraints such as maintaining market share.

As scientists and engineers become more adept at applying prediction and optimization to everyday problems, they are expanding the art of the possible, optimizing everything from our personal health to the houses and cities we live in.

Someone using Google’s self-driving car is completely unaware of the hundreds (if not thousands) of models and the petabytes of data that make it work.

Google realized that the objective was to show the most relevant search result; for other companies, it might be increasing profit, improving the customer experience, finding the best path for a robot, or balancing the load in a data center.

I think of this as a complicated machine (full-system) where the curtain is withdrawn and you get to model each significant part of the machine under controlled experiments and then simulate the interactions.

Great predictive modeling is an important part of the solution, but it no longer stands on its own; as products become more sophisticated, it disappears into the plumbing.

In the future, we hope to see optimization taught in business schools as well as in statistics departments.

Engineers start by defining a clear

objective : They want a car to drive safely from point A to point B without human intervention.

While their models were good at finding relevant websites, the answer the user was most interested in was often buried on page 100 of the search results.

Once we have specified the goal, the second step is to specify what inputs of the system we can control, the

levers we can pull to influence the final outcome.

Once we have specified the goal, the second step is to specify what inputs of the system we can control, the

levers we can pull to influence the final outcome.

In Google’s case, they could control the ranking of the search results.

As predictive modeling and optimization become more vital to a wide variety of activities, look out for the engineers to disrupt industries that wouldn’t immediately appear to be in the data business.

Then, Google came along and transformed online search by beginning with a simple question: What is the user’s main objective in typing in a search query?

Note here the different levels: models of individual components, tied together in a simulation given a set of inputs, iterated through over different input sets in a search optimizer.”

Optimizing for an actionable outcome over the right predictive models can be a company’s most important strategic decision.

They also considered inputs outside of their control, like competitors’ strategies, macroeconomic conditions, natural disasters, and customer “stickiness.”

For example, resistance in the electrical system produces heat, which needs to be included as an input for the thermal diffusion and cooling model.

Only after these first three steps do we begin thinking about building the predictive

models .

A purchase sequence causality model can be used to identify key “entry products.”

A company like Amazon represents every purchase that has ever been made as a giant sparse matrix, with customers as the rows and products as the columns.

One way to escape the recommendation bubble would be to build a

Modeler containing two models for purchase probabilities, conditional on seeing or not seeing a recommendation.

There are many different optimization techniques to choose from (see see sidebar, below ), but it is a well-understood field with robust and accessible solutions.

Instead of the femme-bot voice of the GPS unit telling us which route to take and where to turn, what would it take to build a car that would make those decisions by itself?

Brian Ripley’s seminal book on pattern recognition gives credit for many ideas and techniques to largely forgotten engineering papers from the 1970s.

This curve moves from almost certain acceptance at very low prices to almost never at high prices.

It is easy to stumble into the trap of thinking that since data exists somewhere abstract, on a spreadsheet or in the cloud, that data products are just abstract algorithms.

Amazon’s recommendation engine is probably the best one out there, but it’s easy to get it to show its warts.

Here is a screenshot of the “Customers Who Bought This Item Also Bought” feed on Amazon from a search for the latest book in Terry Pratchett’s “ Discworld series :”

All of the recommendations are for other books in the same series, but it’s a good assumption that a customer who searched for “Terry Pratchett” is already aware of these books.

This has improved emergency evacuation procedures for subway stations and reduced the danger of crowd stampedes and trampling during sporting events.

It is easy to stumble into the trap of thinking that since data exists somewhere abstract, on a spreadsheet or in the cloud, that data products are just abstract algorithms.

The profit for a very low price will be in the red by the value of expected claims in the first year, plus any overhead for acquiring and servicing the new customer.

Once they have the data in this format, data scientists apply some form of collaborative filtering to “fill in the matrix.”

But these products are still just making predictions, rather than asking what action they want someone to take as a result of a prediction.

She cut through the chaff of the obvious to make a recommendation that will send the customer home with a new book, and returning to Strand again and again in the future.

We call it the

Drivetrain Approach , inspired by the emerging field of self-driving vehicles.

Insurers have centuries of experience in prediction, but as recently as 10 years ago, the insurance companies often failed to make optimal business decisions about what price to charge each new customer.

But these products are still just making predictions, rather than asking what action they want someone to take as a result of a prediction.

Many optimization procedures are iterative; they can be thought of as taking a small step, checking our elevation and then taking another small uphill step until we reach a point from which there is no direction in which we can climb any higher.

Someone using Google’s self-driving car is completely unaware of the hundreds (if not thousands) of models and the petabytes of data that make it work.

We are entering the era of data as drivetrain, where we use data not just to generate more data (in the form of predictions), but use data to produce actionable outcomes.

One of the authors of this paper was explaining an iterative optimization technique, and the host says, “So, in a sense Jeremy, your approach was like that of doing a startup, which is just get something out there and iterate and iterate and iterate.”

Great predictive modeling is an important part of the solution, but it no longer stands on its own; as products become more sophisticated, it disappears into the plumbing.

]

ODG identified which

levers the insurance company could control: what price to charge each customer, what types of accidents to cover, how much to spend on marketing and customer service, and how to react to their competitors’ pricing decisions.

There are plenty of cool challenges in building these models, but by themselves, they do not take us to our destination.

The Modeler takes the raw data and converts it into slightly more refined predicted data.

There may be some unexpected recommendations on pages 2 through 14 of the feed, but how many customers are going to bother clicking through?

Full video from that session is embedded below:

Related:

© 2012, O'Reilly Media, Inc.

(800) 889-8969 or (707) 827-7019 Monday-Friday 7:30am-5pm PT

All trademarks and registered trademarks appearing on oreilly.com are the property of their respective owners.

[ Note: Co-author Jeremy Howard founded ODG.

ODG’s competitors use different techniques to find an optimal price, but they are shipping the same over-all data product.

Full video from that session is embedded below:

Related:

© 2012, O'Reilly Media, Inc.

(800) 889-8969 or (707) 827-7019 Monday-Friday 7:30am-5pm PT

All trademarks and registered trademarks appearing on oreilly.com are the property of their respective owners.

ODG’s competitors use different techniques to find an optimal price, but they are shipping the same over-all data product.

We could just build a simple model of distance / speed-limit to predict arrival time with little more than a ruler and a road map.

These products range from weather forecasting to recommendation engines to services that predict airline flight times more accurately than the airline itself.

Someone using Google’s self-driving car is completely unaware of the hundreds (if not thousands) of models and the petabytes of data that make it work.

Google realized that the objective was to show the most relevant search result; for other companies, it might be increasing profit, improving the customer experience, finding the best path for a robot, or balancing the load in a data center.

These firms have plenty of experience building models of each of the components and systems in their final product, whether they’re building a server farm or a fighter jet.

What choice are we actually helping him or her make?

So, we would like to conclude by showing you how objective-based data products are already a part of the tangible world.

On Amazon, the top results for a similar query leads to another book by Toni Morrison and several books by well-known female authors of color.

Insurers have centuries of experience in prediction, but as recently as 10 years ago, the insurance companies often failed to make optimal business decisions about what price to charge each new customer.

Instead of being data driven, we can now let the data drive us.

These days, it is trivial to use some type of heuristic search algorithm to predict the drive times along various routes (a

Simulator ) and then pick the shortest one (an

Optimizer ) subject to constraints like avoiding bridge tolls or maximizing gas mileage.

The objective is to escape a recommendation filter bubble , a term which was originally coined by Eli Pariser to describe the tendency of personalized news feeds to only display articles that are blandly popular or further confirm the readers’ existing biases.

Join the Data Revolution.

What we would really like to do is emulate the experience of Mark Johnson, CEO of Zite , who gave a perfect example of what a customer’s recommendation experience should be like in a recent TOC talk .

Only after these first three steps do we begin thinking about building the predictive

models .

From there, they developed an optimized pricing process that added hundreds of millions of dollars to the insurers’ bottom lines.

These products range from weather forecasting to recommendation engines to services that predict airline flight times more accurately than the airline itself.

Let’s look at how we could apply this process to another industry: marketing.