The post New August Publications appeared first on Jitendra Subramanyam.

]]>**Data & Analytics Strategy and Planning:**

**5 Steps to Get Started with Machine Learning**

https://www.gartner.com/doc/3953653

Eager to get started with ML but afraid it will be too technically difficult, expensive, or time consuming? Click here to learn the 5 steps D&A teams from Micron, Iron Mountain, and Avon used to get started with ML.

**Data & Analytics Quality and Ethics:**

**Human Controls for AI Dangers (SignatureValue Bank*)**

https://www.gartner.com/doc/3947537

Rather than guarding against AI-based attacked, D&A leaders should collaborate with security leaders to guard against the threats internal AI applications cause. Find out how SignatureValue Bank did so here.

**On Demand Problem Solving Teams (McDonalds)**

https://www.gartner.com/doc/3947270

Are you frustrated by onerous data governance practices at your organization? Learn how McDonald’s uses rapid response “sand dune teams” to efficiently make data governance decisions.

**Exclusion-Based Data Sharing Rights (FirstHarbor*)**

https://www.gartner.com/doc/3953614

Do data sharing requests at your organization get bogged down because dozens of stakeholders have to approve them? Find out how FirstHarbor quickly determines which stakeholders should be excluded from reviewing data sharing requests.

**Data & Analytics Talent:**

**Workforcewide Analytics Capability Development (Intel)**

https://www.gartner.com/doc/3955817

Click here to learn how Intel’s Financial Shared Services Center used internal data science talent to increase data literacy across the organization through peer-led education and a community of practice.

The post New August Publications appeared first on Jitendra Subramanyam.

]]>The post Data-Driven Decision Making appeared first on Jitendra Subramanyam.

]]>You are in charge of Group 4. Given that your group’s renewal rate is lower than that of Groups 1, 2, and 3, should you take action to fix customer renewal in your group? Should you find out what’s working, especially in Group 1, and strive to emulate that?

To answer this question, you first have to know the extent to which these renewal rates can vary from group to group without there being any real difference in these rates. Why would the results vary like this? For one of two reasons (or both).

Before we get into the details, here’s the **key takeaway**: *Before* you make any data-driven decisions, figure out how much variation you *should *expect to see within the groups based on these two reasons alone.

Let’s get into the two reasons for variation in data of the kind we see in Figure 1.

First, consider the entire population that consists of all of the groups put together. We can think of it like the diagram below.

The green plus signs are customers who will renew at the end of the year. The red minus signs are customers who won’t renew at the end of the year. The population of all customers can be grouped in many ways. Here are just two of the many possibilities.

You can see why, just by grouping customers differently, the renewal rate of the groups can change. This kind of variation in the percentage renewal rate within a group is called *sampling variation.*

Now, there might be a very good reason why you group customers the way you do. It might indeed be the case that a customer in Group 4 cannot be conceived to be part of Group 2, and vice versa. But sometimes, group affiliation is fluid. A customer in Group 4 could well be classified as falling into Group 2 or vice versa. This can happen when the groups themselves are not conceptually watertight or there is an issue with data quality that makes it difficult to definitively classify customers into one group or another.

In such situations, just seeing a difference in percentage renewal rates between the groups is not enough to drive action. We must first determine if the difference lies within the variation we *should *expect. If it does, then there’s no reason to take action; if the differences are greater than what should be expected due to sampling variation, then we are seeing a real difference between the groups.

To determine the variation we should expect to see, we can do a bit of simulation. Imagine you have a *population *of clients as shown in Figure 2. For this population as a whole, we’ll set the renewal rate at 80% — in other words 80% of the total number of customers renew at year end. From this population we’ll choose a random sample of 300 customers (*without* replacement, in case you’re wondering). We’ll calculate the renewal rate of this sample. And we’ll repeat this 100 times to get a sense of how the renewal rate varies from one sample to another.

Here’s the variation we see when run the simulation.

Just by taking different samples from the same population, the renewal rate can vary from a high of roughly 86% to a low of roughly 75%. Remember, this is the exact same population with the exact same renewal rate of 80% overall. But simply because we sample different groups of 300 customers from this population, the sample renewal rates range over a difference of 11 percentage points!

Figure 5 depicts the same point using Group 4 as an example. Sampling variation alone will result in Group 4’s percentage renewal rate to vary anywhere between 86% and 75%.

In other words, there is no real difference between a percentage renewal rate of 86% and 75% — any value in this range is *equally likely *due to the inherent variation in the way Group 4 (or any other Group for that matter) is composed.

Except for Group 5, sampling variation alone accounts for the range of values we see in Figure 1. There is really no difference in renewal rate between Group 1 and Group 4 although it appears so at first glance. It would be misguided to take action based on Figure 1 to fix Group 4’s renewal “problems”. However, the customers in Group 5 have a renewal rate of 72% which falls below the range we see; it makes sense to look into the customers in this group and take appropriate action to improve their renewal rates.

Variation in the results seen in Figure 1 could be for a second reason as well. Let’s understand this second reason for variation.

Put yourself in a customer’s shoes. During the course of a year, a customer experiences various events that may influence his or her propensity to renew. Some of these events positively influence the renewal decision; for example, having a great interaction with customer service or saving critical time by using the product. But many of the events during the course of the year might negatively influence the renewal decision. The customer’s role may change making the product no longer useful, or they may switch employers. Or they may have a bad experience with the product.

How many positive and negative events a customer might encounter during the course of a year is hard to predict. But this kind of customer renewal trajectory can be modeled based on some assumptions. We’ll use three such models to try and capture the customers’ trajectories. The point is that we can’t say much about the data without having an understanding of how the data was *generated*. That’s what a model aims to do. Let’s look at our first trajectory variation model.

A Brownian Process is a model that captures the many vagaries of a customer’s renewal journey. The customer starts the year at a neutral state, say at zero. Through the course of the year the customer experiences positive and negative events. A positive event pushed them in the positive direction along the y axis while a negative event pushes them in the negative direction along the y axis (see Figure 6 below).

The Brownian model works by setting the number of events a customer experiences and uses a Normal distribution of positive and negative events from which the particular events are randomly selected.

Once again, we take 300 customers and simulate their trajectories over the course of a year. We find the renewal rate for this group. We then repeat this again until we’ve run 100 simulations. The results are captured in Figure 7 below.

And once again renewal percentages range from 86% to 75% showing that there is no real difference between the renewal rates between the first 4 groups shown in Figure 1. There is no reason to take action to fix Group 4’s renewal rate based on what we know about the range of variation in renewal rates.

In a Markov process, a customer starts out the renewal year in a particular state. They’ve just signed and are excited by the product (for the most part). Let’s call this the “High Value” state. During the course of the year the customer experiences various events. Each event leads to a transition from one state to another *based only on their current state. *In a Markov process these transitions are defined by a *transition matrix*.

If a customer’s current state is High Value, no matter what event they experience, they have a probability of 0.6 of staying in the High Value state. Similarly, if a customer is in a Medium Value state, they have a probability of 0.1 of transitioning to a Low Value state no matter what event they experience. And once they are in a Low Value state, the probability of transitioning to a Medium Value state is 0.3.

We can set the transition matrix to anything reasonable. But note that the Markov process model assumes that a customer’s current state is more important than the type of event they experience, positive or negative. In this aspect, the Markov process model is quite different from the Brownian process model where the customer is swayed in a positive or negative direction by the positive or negative valence of each event they experience.

Once again we simulate 100 instances of 300 customers moving through their renewal journeys. Each customer has 24 interactions — or actually, 24 transitions from one value state to another. At the start of the year 40% of customers are in the High Value state, 50% in the Medium Value state and 10% are in the Low Value state. The renewal rates at the end of the year for 100 simulations is shown in Figure 9.

In Figure 9 we again see variation in renewal rates from 86% to 75% due to the multiple ways in which customer renewal journeys unfold. This variation in customer renewal rates is what we should expect. Going back to Figure 1 we again see that that there’s no real difference between Groups 1 through 4 even though the reported renewal rates differ.

While the Brownian and Markov process models capture some of the aspects of a customer’s renewal trajectory, they fall short as models in some ways. Perhaps it’s not always appropriate for events to come from a Normal distribution (as is the case in our Brownian process model). Or maybe the customer’s mindset is not as important as the Markov process model makes it out to be. To get around these problems we can make assumptions that better fit the customer renewal trajectory.

For our customized random process model we’ll once again use 300 customers. During the renewal trajectory a customer experiences six types of events:

- Positive High Impact (Prob = 0.1, Impact = +5)
- Positive Medium Impact (Prob = 0.1, Impact = +2
- Positive Low Impact (Prob = 0.1, Impact = +1)
- Negative Low Impact (Prob = 0.24, Impact = -1)
- Negative Medium Impact (Prob = 0.23, Impact = -2)
- Negative High Impact (Prob = 0.23, Impact = -5)

Each customer is randomly assigned the number of events they experience during the course of the renewal year. The distribution of the types of events is set as described above — 30% of the events are positive while 70% of the events are negative. A customer starts off the year at zero and is pushed farther from or closer to this starting state based on the number and types of events they experience. At the end of the year they are either above or below a set threshold. If they end up above the threshold, they renew, otherwise they don’t.

Here’s the variation in renewal rates when we use the customized random process model and simulate the journeys of 300 clients.

You won’t be surprised to see the variation of renewal rates in Figure 10.

Without knowing how much variation there is between groups — variation due to sampling, trajectory variation or both — it doesn’t make sense to read reported differences in dashboard visualizations as real differences. It makes no sense to act on differences that are caused simply by the types of variation we’ve seen. To make effective data-driven decisions (rather than misguided ones) figure out if the differences exceed what you should expect to see. To figure out how much variation you should expect to see, get a better understanding of the process by which the data was generated.

The post Data-Driven Decision Making appeared first on Jitendra Subramanyam.

]]>The post When Machine Learning Prediction Excels appeared first on Jitendra Subramanyam.

]]>In the previous post, Prediction Models: Traditional versus Machine Learning, we looked at 3 kinds of prediction models and clarified the difference between traditional and machine learning models for prediction. In this post we’ll see that machine learning prediction models excel in conditions in which other prediction models suffer.

Let’s look at how each type of prediction model (Traditional Statistical, Traditional Rules-Based, and Machine Learning) satisfies some key aspects that characterize them. Specifically, we’ll look at the the following:

- What’s the method for determining the model’s optimal parameter values?
- Does the method for determining the model’s optimal parameter values need to be programmed by humans?
- Can the method for determining the model’s optimal parameter values handle data sets with a large number of rows and/or columns?
- Can the method for determining the model’s optimal parameter values

The table below captures the differences between the prediction models.

Recall from the previous post that the optimal parameters are the values that enable a model to make specific predictions. When the parameters of a model are not set, the model is a formula that is incomplete — the form of it is complete, but there’s not enough information to calculate a number based on the formula. Here’s an example:

a * 16 + b * 12 = ?

We don’t know the answer to this equation until we know the specific values of the parameters a and b.

While the traditional statistical models require a lot of creativity (and statistical sophistication) to determine the parameters, the most common algorithm that machine learning models use for this purpose is called *gradient descent*. This algorithm is conceptually simple and I’ll explain it in a future post.

The gradient descent algorithm proceeds iteratively and ultimately discovers the optimal parameter values. But don’t confuse an algorithm which is programmed to discover with an algorithm that “writes itself”. Sometimes you hear that machine learning is all to do with computers writing their own code; as row 2 of the table above points out, this claim is fantasy.

The last 3 rows of the table above outline the conditions in which machine learning models thrive while traditional models struggle. The machine learning prediction approach is particularly suited to data sets that:

- Have a large number of columns (each data point has a large number of attributes)
- Have a combination of categorical, numerical, and textual (or image, audio, video) data

It pays to try machine learning prediction models when you face these conditions. Especially if other methods haven’t been able to make reliable predictions and there is a lot of business value gained in beating the existing prediction benchmark.

Bonus points if the marginal business utility of beating the benchmark prediction accuracy is also high. This means that even small improvements to the existing benchmark are valuable.

The post When Machine Learning Prediction Excels appeared first on Jitendra Subramanyam.

]]>The post Prediction Models: Traditional versus Machine Learning appeared first on Jitendra Subramanyam.

]]>Machine learning models are constructed differently from traditional quantitative models.

In the first type of traditional prediction model, the input data set along with statistical assumptions and calculations determine the prediction algorithm. The input data set is analyzed (or “fitted to the data”) using statistical techniques. The prediction algorithm that is the one best suited to describing the data as determined by the statistical analysis.

The second type of traditional prediction model uses an explicit set of rules (e.g., if X then Y) to transform the inputs into a prediction. Instead of the prediction algorithm being “discovered” through statistical calculations, these rules are usually ones that are known by experts in the prediction domain (e.g., the medical knowledge physicians have in diagnosing/predicting a disease).

In contrast to the traditional quant prediction models, machine learning prediction models are developed in two steps.

In the first step, a machine learning model is trained. In this training step the input data, the historical results associated with these inputs, *and *a training algorithm are used to iteratively arrive at the prediction algorithm.

At the end of the first step, the model is “trained” and is now ready to make predictions.

In the second step, the prediction step, the trained machine learning model uses the prediction algorithm arrived at in the training step to transform new inputs into predictions.

You may have noticed that I’ve used “model”, and “algorithm”. Is there a difference? This gets us to some terminology that’s useful to have in mind when you read about machine learning.

A *model *is a mathematical expression that transforms inputs into outputs. The model by itself cannot be used to calculate a result. To do so requires fixing the model’s *parameter *values. In traditional approaches, the parameter values are fixed based on statistical calculations. In machine learning the parameter values are fixed in the process of training the model. Without its parameter values specified, a model is a structure, a shell, a high-level directive for transforming inputs into outputs. With the parameter values specified, the model can be used to generate specific predictions.

An *algorithm *is a recipe — a clearly defined step by step by step process — for turning a set of inputs into an output.

A *prediction algorithm *is the specific mathematical expression that results once the parameter values of the model are fixed. In machine learning these parameter values are fixed during the iterative training process which itself proceeds by using an algorithm. The most common of these training algorithms is called *gradient descent.*

The post Prediction Models: Traditional versus Machine Learning appeared first on Jitendra Subramanyam.

]]>The post How Long Should You Look Before You Leap? appeared first on Jitendra Subramanyam.

]]>In his fascinating book, *A Man For All Markets*, Edward Thorp trots out a problem called the “secretary problem” and breezily gives us the solution (pp.263-4).

“Assume that you will interview a series of people, from which you will choose one. Further, you must consider them one at a time, and having once rejected someone, you cannot reconsider. The optimal strategy [the one that maximizes the probability of picking the best one] is to wait until you have seen about 37 percent of the prospects, then choose the next one you see who is better than anybody among this first 37 percent that you passed over. If no one is better you are stuck with the last person on the list.”

But why should this be? What’s so special about looking at but not selecting the first 37% of applicants and then leaping to select the next candidate who comes along who’s better than any of the first 37%? You can look up the mathematical solution to the problem. But is there a simpler way to get this result, a more intuitive way to understand why 37%?

There is. It’s through simulation. Let’s set it up.

Suppose we have n elements that are rank ordered from 1, 2, …, n. The elements are jumbled up so the rank ordering is messed up. For instance they are jumbled up as 4, 56, 73, 1, …, n, 32. We can only see the elements one at a time. The only choice we can make is to either select or reject an element when it is presented to us.

If the objective of the game is to maximize our chances of selecting the best element (which we’ll designate as element 1, second best as element 2, and so on), what strategy should we adopt? That is the secretary problem. It might seem like there’s no good strategy. But as Thorp pointed out, there is.

The winning strategy is to adopt the Look-Then-Leap algorithm *where you switch from look to leap once you’ve looked at 37% of the total number of items*.

Here are the steps of the look-then-leap algorithm:

- Look at and reject the first k items where k < n
- Go to the next item in the list – if this is the last item, choose it and STOP
- Check if the item is better than any of the k items
- If yes, then choose the item and STOP
- If no, then GO BACK TO STEP 2

The secretary problem has the following characteristics:

- We have a finite list of things to choose from
- These things appear before us one by one
- They can appear in any order
- The items are ranked — the higher the rank, the better the item
- We don’t want to leap too early or we’ll risk missing a better candidate who is yet to appear
- We don’t want to leap too late or we’ll risk rejecting the best candidate

There are many problems that fit these simple constraints. In their book *Algorithms To Live By*, Brian Christian and Tom Griffiths point to hunting for an apartment in San Francisco a hot market or finding a mate as examples. My colleague Kevin Gabbard generated this list of similar problems.

In general, if we had 3 candidates (the best ranked 1, the second best ranked 2 and so on), they could appear for interviews in any one of 6 orders: [1, 2, 3], [1, 3, 2], [2, 1, 3], [2, 3, 1], [3, 1, 2], [3, 2, 1]. Let’s apply the Look-Before-You-Leap algorithm by looking and always rejecting the first one and then leaping for the next best one. Since we have 3 items, looking at the first one and leaping thereafter means leaping after 33% of the candidates have been observed.

If we look at the first candidate and reject them and then leap for the better candidate, 50% of the time we end up with the best candidate. While not great, this is significantly better than choosing at random which yields the best candidate only 33% of the time.

And so on with 4, 5, and any number of candidates. But as the number of candidates grows, the number of permutations grows very quickly. While there are 6 ways in which 3 distinct things can be permuted, there are more than 3.6 *million *ways in which 10 distinct things can be permuted. So in order to simulate the situation of selecting from a pool of 100 candidates, we can’t possibly check all permutations in which they appear.

Rather, we’ll start with the candidates in order 1, 2, 3, …, 100 as one of the possibilities. We’ll then scramble up the order randomly many thousands of times to simulate a the experiment of choosing a candidate based on the Look-then-Leap algorithm. Each time we pick a candidate based on the look-then-leap algorithm and then at the end of the run of simulated trials we check to see what proportion of the picks are picks of the best candidate.

Here are the results when we simulate 200,000 trials for a pool of 10 candidates and then for a pool of 100 candidates.

With 10 candidates to choose from in total, we were told that the best strategy is to look at 3 or 4 candidates first, reject them and then choose the next candidate who’s better than the ones initially rejected.

The simulation indeed validates looking at (roughly) 37% of the pool before leaping. The simulation also reveals three other curious facts:

- The probability of choosing the best candidate is also around 37%. So you look at 37% percent of the pool before you leap; and once you leap, you have a 37% chance of selecting the best candidate. This is a neat coincidence!
- The probability of choosing the best candidate is
*significantly higher*than choosing the second or third best candidate. This is quite counter-intuitive and usually not the way things tend to be. - And finally, it doesn’t matter if there are 10 candidates or 100 (or any number): you should always leap after you’ve looked at 37% of the pool in order to maximize the probability of selecting the best candidate.

We can check this third point by simulating the look-then-leap algorithm for a pool of 100 candidates.

In their book *Algorithms to Live By*, Brian Christian and Tom Griffiths underscore the miraculous nature of this third point (p.15):

“If we were hiring at random, …then in a pool of hundred applicants we’d have a 1% chance of success [of hiring the best applicant],

in a pool of a million applicants we’d have a 0.0001% chance. Yet remarkably, the math of the secretary problem doesn’t change. If you’re stopping optimally, your chance of finding the single best applicant in a pool of hundred is 37%. And in a pool of a million, believe it or not, your chance is still 37%. Thus the bigger the applicant pool gets, the more valuable knowing the optimal algorithm becomes. It’s true that you’re unlikely to find the needle the majority of the time, but optimal stopping [i.e., using the look-then-leap algorithm] is your best defense against the haystack, no matter how large.”

The post How Long Should You Look Before You Leap? appeared first on Jitendra Subramanyam.

]]>The post Chief Data and Analytics Officer Research Publication List appeared first on Jitendra Subramanyam.

]]>__Jitendra Subramanyam __ __https://www.linkedin.com/in/jsubramanyam/__

__Ben Hertzberg ____https://www.linkedin.com/in/benjaminhertzberg/__

__Farhod Yuldashev ____https://www.linkedin.com/in/fyuldash/__

__Ethan Green__ __https://www.linkedin.com/in/ethanfgreen/__

Our research terrain covers:

- Business Value of Data & Analytics
- Data & Analytics Strategy and Planning
- Data & Analytics Quality and Ethics
- Data & Analytics Talent

**5 Steps to Get Started with Machine Learning***NEW!!*

Eager to get started with ML but afraid it will be too technically difficult, expensive, or time consuming? Click here to learn the 5 steps D&A teams from Micron, Iron Mountain, and Avon used to get started with ML.**How to Reveal the Business Value of Imperfect Data with AI (Avon)**

Imperfect data is worthless for business intelligence. But it can create business value, if organizations switch from BI to advanced analytics. Find out how Avon did so here.**Data and Analytics Value Creation: Key Obstacles and How to Overcome Them**

Learn what Chief Data & Analytics Officers polled during the 2019 D&A Summits in London and Orlando believe enables their organizations to create business value with data & analytics.**Peer-Based Analytics Learning (ABB)**

Frustrated with low analytics use in your organization? Take a lesson from ABB’s audit function: they use peer-led case studies to give auditors hands on experience in how analytics can improve audits.**Machine Learning Literacy for Business Partners**

Do you have data scientists mired in dashboard creation? Or do they develop cool products that don’t meet business needs? Find out how Micron improves communication between data science teams and business partners with a simple ML literacy course.**Machine Learning Literacy for Business Partners (Micron) Implementation Tool**

Download Micron’s internal ML literacy syllabus here. It includes two case studies business partners can use to experience developing an ML solution on their own.**Analytics Presentation Engagement Framework (NGA)**

Too often, analysis falls on deaf ears, and excellent insights fail to drive value. See how analytics leaders at the National Geospatial-Intelligence Agency create analytics presentations that motivate business users to action.**Opt-Out Decision Engineering to Increase Analytics Use**

Business users often have powerful analytics tools available—but they rarely use them. Data and Analytics leaders can use an opt-out technique to shape the behavior of business users by “nudging” them into using of analytic tools.**Decision-Focused Data Maps (General Mills)**

Do people in your organization spend more time looking for the right data than using it to inform decisions? Find out how General Mills developed an easy-to-understand visual that connects crucial business questions to available data sources.**Simple, Powerful Machine Learning Pilot (Iron Mountain)**

Do worries about expertise and expense keep your organization from piloting ML projects? Find out how two FTEs in Iron Mountain’s A/R team developed a Machine Learning pilot off the side of their desks that decreased time to payment by 40%.**From Data to Prediction (Iron Mountain): Further Details**

Find out the specific steps Iron Mountain used to develop their Accounts Receivable late payment prediction pilot here.**How to Build Momentum for Machine Learning (ML) Initiatives (Iron Mountain)**

D&A leaders need to build on their successes with Machine Learning. Find out how Iron Mountain did so here.

**Continuously Market-Tested Data & Analytics Strategy (UrbanShopping*)**

Creating value from enterprise data requires organizations to make a blinding array of choices. UrbanShopping’s D&A strategy led them to create a D&A sandbox that enabled the rapid market testing of D&A solutions and drove substantial ROI.**IT Score for Data and Analytics**

Uncertain how to get started in D&A? Use Gartner’s IT Score for Data & Analytics to measure D&A maturity across seven objectives and 25 discrete functional activities!**Ignition Guide to Strategic Planning for Data & Analytics**

As a Data & Analytics leader, how do you develop a world class D&A strategy? Tap into the collective wisdom of hundreds of D&A leaders. Here is Gartner’s step-by-step guide, with tools and templates, to help you establish an actionable Data & Analytics strategy.**Analytics Prioritization Principles (Gap Inc.)**

How can Data and Analytics leaders sense, prioritize, and satisfy the critical data and analytics needs of their business users? Gap Inc. provides a model for engaging with business users to determine their data needs and priorities and develop the analytics they need.**Data & Analytics Strategy Workbook**

D&A leaders often struggle to navigate the complexities of the strategic planning process. This workbook outlines the steps involved and provides hands-on tools and templates to create a strategic plan document.**Data & Analytics Strategy Presentation Template**

This template provides D&A leaders with customizable recommended and optional slides to craft an effective D&A strategy presentation.**Data & Analytics Sample Strategy Presentation**

This sample strategy presentation is an illustrative example of how to tie D&A strategy to business strategy and improve organizational decision making through analytics investments.**Planet Architecture: The Role of Data in Platform Strategy (A Conversation with Ian Reynolds and Jitendra Subramanyam)**

Data is crucial to building the business case for new technology platforms. This episode explores how EAs and D&A leaders can maximize the business value of a platform strategy by effectively using available enterprise data.

**Human Controls for AI Dangers (SignatureValue Bank*)***NEW!!*

Rather than guarding against AI-based attacks, D&A leaders should collaborate with security leaders to guard against the threats internal AI applications cause. Find out how SignatureValue Bank did so here.**On Demand Problem Solving Teams (McDonald’s)***NEW!!*

Are you frustrated with onerous data governance practices at your organization? Learn how McDonald’s uses rapid response “sand dune teams” to efficiently make data governance decisions.**Exclusion-Based Data Sharing Rights (FirstHarbor*)***NEW!!*

Do data sharing requests at your organization get bogged down because dozens of stakeholders have to approve them? Find out how FirstHarbor quickly determines which stakeholders should be excluded from reviewing data sharing requests.**Business-Need Driven Data Governance Objectives (FirstHarbor*)**

Does your data governance enable value creation or constrain it? Find out how FirstHarbor narrows the scope of data governance, meeting business needs in data collection, use, and sharing while ensuring compliance and productivity.**Value-Add Data Minimization (Northrop Grumman)**

More data is not always better, because data comes with risk. Find out how Northrop Grumman selects data for minimization and sells business partners on value-adding alternatives.**Ignition Guide to Building a Data and Analytics Governance Program**

Don’t assume that traditional data governance will meet the demands of big data and digitization! Use this guide to establish a governance program that aligns with business priorities and divides strategic and tactical responsibilities.**Data Quality Score (TE Connectivity)**

Does your organization struggle to get buy-in for data quality improvements from your business users? Find out how TE Connectivity used an enterprise-public data quality score to hold business users accountable for data quality.**Dangerous Data: Can’t Live Without It, Can’t Live With It**

Many D&A leaders think data is good and more data is better. But some data can pose serious legal, ethical, and brand risks to organizations. Find out why here.**“Show Don’t Tell” Data Quality Improvement (Citizens Bank)**

Business users often struggle to see the relevance of data quality to their work. Citizens Bank creates business demand for increased data quality by contrasting the insights and reports that could be generated from higher quality data to the current state.**Data & Analytics Operational, Data Quality, and Data Management Metrics**

How do Data and Analytics leaders measure success? Find out here: our collection of real-world metrics spanning operations, business value, data quality, and data management maturity.

**Workforcewide Analytics Capability Development (Intel)**Click here to learn how Intel’s Financial Shared Services Center used internal data science talent to increase data literacy across the organization through peer-led education and a community of practice.*NEW!!*

**Data and Analytics Job Descriptions Library**

From cutting-edge, emerging roles to those that are now standard in D&A teams, here’s where to find job descriptions sourced from peer organizations. Updated quarterly with contributions from Gartner’s entire D&A research community!**Redefining Analysts as Decision Experts (Philips)**

Find out how Philips grew revenues by more than 18 million by aligning its analysts to support specific decision areas rather than a myriad of stakeholder requests.**Creating Business Value with Multidisciplinary Data and Analytics COEs (Omicron)**

Does your organization use a Data and Analytics Center of Excellence (COE)? Are you thinking of setting one up? Learn how Omicron avoided silos and enabled cross-functional collaboration in their COE for Finance D&A.**D&A Organizational Models, Roles, and Responsibilities**

How do Data and Analytics leaders organize their function? What roles should a data and analytics office have? This deck collects practitioner examples of organizational, staffing, and stewardship models and analytics roles.**Analytics Champions Recruitment Guide**

Learn how companies identify internal evangelists for analytics and use them to increase analytics demand across the enterprise.

The post Chief Data and Analytics Officer Research Publication List appeared first on Jitendra Subramanyam.

]]>The post Meet the Chief Data & Analytics Officer Research Team appeared first on Jitendra Subramanyam.

]]>We’re a new Gartner research team. We produce best practice case studies, tools, and templates. The content of our research comes from leading practitioners of Data & Analytics from organizations around the world.

That’s us left to right: Ben (LinkedIn), Ethan (LinkedIn), Farhod (LinkedIn), and, Jitendra (LinkedIn).

- Generating Business Value with Data & Analytics
- Creating Actionable Data & Analytics Strategies
- The Quality and Ethics of Data & Analytics
- Data & Analytics Talent

We have diverse educational backgrounds spanning political science, economics, management, international development, physics, philosophy, statistics, machine learning, and engineering. Together we speak at least 10 languages (Farhod has everyone beat — he speaks 5). We’ve worked in particle physics labs, non-profit organizations, universities, software companies, and state and local governments.

We like reading, hiking, cooking, stand-up comedy shows, traveling, and playing sports – we are hosting our first ping pong world series championship within the team next month.

Follow our journey as we blog about our D&A research.

Next installment: a list of our current research…see you again soon!

The post Meet the Chief Data & Analytics Officer Research Team appeared first on Jitendra Subramanyam.

]]>