# Data-Driven Strategy: Aligning Executive Bias With Customer Value

**Podcast:** HBR IdeaCast
**Published:** 2026-11-10

## Transcript

I'm Adi Ignatius.
I'm Alison Beard, and this is the HBR IdeaCast.
So, Alison, ask any good business leader what their customers value, and they will say they know, right?
It is their job to know.
They have the data, they're checking in all the time on the front lines with their customers, and they have years of experience in the industry.
Right, and presumably good leaders also act on that information.
They make a plan to better serve their customers.
Absolutely, and yet, according to Vikas Smithall, a professor of marketing at the Jones Graduate School of Business at Rice University, companies often collect the data but then build a plan instead that is based on their intuition.
So they miss what matters most to their customers and they miss what to prioritize to create the most value for the company.
So I do love intuition, but if organizations are collecting the data and the answer is in there, then they really should follow through on what they've learned.
Yeah, I mean, instead, people are overriding what the data says pretty clearly.
which, of course, leaves a lot of money on the table.
So Mindel spent seven years studying 72 businesses, and he has lots of advice on how companies can get better at understanding the data that they have at hand.
He's the co-author of the HBR article, Do You Know What Your Customers Value?
Here's our conversation.
I have a lot that I want to ask you.
I want to start, though, maybe you can set this up for our listeners.
You and your colleagues recently wrapped up a seven-year study.
They compared what executives thought customers valued versus what customers actually valued.
Talk briefly about what you found.
So the main finding is that executives more or less are always wrong in terms of predicting what creates value for customers.
They're good at telling you what customers say they want, which is different than what creates value for customers.
Okay, that is a good tease.
All right.
These are serious companies.
They do a lot right.
They're collecting data.
They're checking in with their customers.
They're being led by executives with a lot of experience.
What happened?
What was missing from the process?
Where did it go wrong?
Yeah.
So the missing link in the process is what we call the DAW's insight, right?
The idea is that when you just ask people what you would want, which is...
Generally, the process we use, we take customers out, you know, set up focus groups.
We talk to them, basically say like, you know, what are your pain points?
What would you like us to do better, et cetera, et cetera.
At that point, customers can tell you what they want, but customers often cannot tell you what they value.
And my example of that is, you know, in apartment buildings.
They ask customers typically, what would you want?
And many customers say, we want a gym in the apartment.
Yet, fewer than 20% of the customers actually use the gym.
If you do the research correctly, what you find is what creates more value for the customers is typically not the gym, but the building administrators to fix any problems, provide any service if there's like broken appliances and stuff like that.
So this is so interesting.
I mean, this is almost, you know, sacrilege.
I mean, I talk to leaders all the time and they say the secret is get close to your customers, get to the front line, talk to the people who actually buy your product.
But I mean, you're suggesting that this sort of vivid customer interaction can actually distort strategy.
It typically does because exactly what you say, customers can recall things that are vivid to them.
Customers can tell you things that are salient to them.
Customers can tell you things that just happened recently.
But many of these things are not those things that create value for the customer, right?
Often things that create value for the customer, many of these are just very small things that need to be happening on a routine basis.
When I looked at the data for many of the companies, when you actually see things that create the biggest lift in value for the customers are so minutiae and so small that companies often don't pay attention to them, right?
Like, so the ITI example we talk in the paper is basically a weekly update for the customer, which is a small email with three bullet points.
This is what's going on with your project.
So really, it's not a matter of CEOs needing to listen differently, but really to not put much faith in these conversations and to trust the data.
I mean, that's what you're saying, right?
Yes.
Yes, absolutely.
Like, you know, it's not that you don't need to listen to the customers.
You should, you can, but that is not a good basis to create your entire strategy.
So before we get to the data, I mean, this is also about blind spots.
Can you talk about how we have either consciously or subconsciously in attachment?
to our own perspectives, our own experience, let alone biases.
I mean, we tell ourselves in strategy sessions all the time, don't do that.
But, you know, how do we actually not do that?
Because you and your colleagues, your research seems to suggest that we're doing that a lot.
Yes, the way you don't do it is actually just don't do it.
So you think, right, like these customer conversations.
There are many reasons to do it, right?
Like I just want to show my appreciation to the customer.
I want the customer to feel valued.
I want the customer to feel that they have been able to open up to me and lay out urgent things that need to be addressed.
The customer telling you my delivery has not shown up and you can be reactive to that.
What our point is, all of those things are fine, but creating a strategy based on that that is going to guide your company for the long run is a mistake.
So explain your method to listeners, to me, to people who are not statisticians.
Explain what you're looking at and how the process works.
The basic idea, and it's been around, like in 1979, Robin Dawes published an article.
And the main point in that article was that if you have something that you want to predict.
A statistical linear model can always predict that better than the expert's judgment, right?
So you may go and talk to five customers and say, what do you want or what creates value for you?
You know, somebody would say like, we want a safer product.
Somebody would say we want lower price.
The problem is when we come back, you know, you may have a bias, Adi, because for some...
reason like you're more into safety.
Maybe you're the chief safety officer of a company.
I'm the chief sales officer.
I may have a bias about lower pricing.
We are going to push that thing through because of our blind spots or preconceived notions.
But if you were to go in, give a structured assessment to all 10 customers, and then from there, you could statistically figure out across all 10 customers, did safety create more value or did lower price create more value?
And that's not dependent on your judgment or my judgment.
That's dependent simply based on what the statistics are telling us.
And that's what the strategy should be based on.
So talk, let's say, one more level of detail, how the data is showing you exactly what customers truly value.
Perfect.
So the way we would analyze the data very rudimentarily, you know, there's many bells and whistles.
On the x-axis, you could simply plot.
the extent to which safety creates value.
And on the y-axis, you could plot the extent to which there is overall value in this offering.
And if that slope is high, that means if you increase safety, you also get more value.
You could draw a similar plot, let's say for pricing.
And what you find is, as people find more value in pricing, they may not find more value in your offering.
Pricing should not be the focus of your company's strategy.
It should be safety.
So it's in some ways like a very basic correlation between the value that safety brings and the total value customers experience from your offer.
You talked about safety.
I sort of wonder where customers would take certain things for granted, right?
They would take safety for granted and might indicate that they put a value elsewhere.
Is it possible to tease apart?
You know, what are they taking for granted versus, you know, what do they truly need and might not necessarily get if they don't express that preference?
I'll answer it slightly differently.
You know, so customers have baked in expectations of so many things, right?
So you can easily tease apart if your expectations are met and you perform above expectations, you may get no additional lift in value.
Safety is one of those things.
Right.
Everybody expects you to be safe, like if you're an airline.
So if it's just safe, you get no additional lift.
Take the exact opposite.
It could be something related to in-flight service.
Right.
But if you get extraordinarily good service, you may actually get an additional lift.
So the idea is that meeting expectations, not meeting expectations versus exceeding expectations can all have an asymmetric effect on value.
you should statistically incorporate all of that.
I'll give you an example.
One of the companies I did this work for, they were looking at on-time delivery.
What they found is that delivering a day ahead of the intended date did not create any additional value for customers.
But being a day late created a substantial decrease in value.
So they put some processes in place because delivery ahead of schedule was very expensive for them.
So they created this whole idea of on time to need date, where they delivered on time to the need date, but not ahead of it.
And that really reduced their expenses.
So what's your advice for companies to respond to all of this?
So on the one hand, gather data properly, pay attention to it.
try to figure out what truly is driving value.
But, you know, in terms of generating the data, having an approach, a culture that will gather the data, will respect the data, how do you even start?
Yeah.
To think of strategy planning and strategy formulation as a very science-driven process rather than a process of subjective judgments.
So often, I'm sure you've seen this, when we say we are going to do a strategy retreat, a lot of it starts by people getting into a room and just like brainstorming, what do we think customers want?
Right?
And a lot of that is relied on like, oh, you know, I was talking to our lead sales guy the other day and he said there was this one customer who wanted that, right?
Then somebody else will dig into their memory bank and pull up another thing.
And then the third person will dig into their memory bank.
thing I would say is less reliance on memory or simple anecdotes to create a strategy.
And so we talk about that in the paper, right?
Like salience is the culprit here.
And then the second thing, just because something created value for your competitors doesn't mean you should blindly replicate it and assume that it also creates value for you.
So those are the starting steps.
So one approach is, okay, if I have a hunch, if I talk to a customer and I get a sense of what we need to do, you create an experiment, right?
You don't necessarily redirect the company in a new direction, but you create an experiment and then get some data.
I think what you're saying is the data can tell you before you even need to go into the experimentation stage what the most likely successful experiment will be.
Is that right?
That's exactly correct.
In fact, the way I would think about this is when you start listening like that, it can create all the conditions in the experiment, right?
So you heard safety, I heard pricing, somebody else heard quality, another person heard on-time delivery, another person heard something else.
So all of these are now testable variables.
And our goal is now to use science to rank order these variables in terms of which creates the most value.
So this listening exercise, trying to see what the environment is presenting, it's an ideation in terms of what are all the levers that we can actually influence, right?
And the science can help you to rank order the levers.
And then the implementation is honing on those levers that create the most value.
So at some point in a conversation like this, inevitably we have to have the Steve Jobs discussion, right?
So, you know, Steve Jobs is our model and sometimes our self-image of, you know, forget the focus groups, forget the data.
People only know what they have experienced.
They can't possibly point the way to the future.
And at a certain point, you need your insights.
You need whatever sense you have of the market that's based on your experience that will not show up in the data.
So how do you handle that?
In my life, there are like two or three CEOs who can figure out what creates value without doing a statistical model.
Steve Jobs was one.
The other, I think, is Warren Buffett.
And we are ordinary beings.
We are neither of them.
Well, that's certainly humble.
So I think if we just peel back the onion in terms of what Steve Jobs was really doing.
He had an uncanny knack from separating what a lot of people wanted versus what actually created value for them.
And this is a very crucial point, right?
Like here, we are using statistics to identify what creates value for the customers.
But if you had just asked customers what you want, you would have been misled.
And that's exactly what he did without using statistical models, though.
So there is the human element, the personal element of strategy that you can't ignore.
And every strategic initiative has an executive sponsor, has a human, has a person.
When your analysis, let's say your approach, seems to indicate that a powerful leader's signature program adds little customer value.
Realistically, how does an enterprise deal with that disconnect?
If you watch the new Star Trek, when they ask, Mr.
Spock, how did you defeat Khan?
And he says, like, with grave difficulty, right?
So what I have seen in organizations is two things.
If you were to just suddenly approach somebody and say, like, Bakas, we don't value your initiative and we need to change it, there's a natural reaction.
Well, this might be political, blah, blah, blah.
But there's a different way to approach it, where if you've done things with science.
then the idea is to have a conversation in terms of where we need to focus on as an organization.
And what I have found is CEOs are very successful if they go to a person not just saying that we want to stop this initiative, but having a different conversation, which is, look, this is one thing that creates even more value than your initiative, Bakas, and we want to have a conversation.
how you can participate in this thing that's even more valuable, right?
And that conversation, because it starts from science, it gets less pushback.
I'll give an example.
So the Swagelog company that's referenced in the HBR article, they were all set to put tons of money into an initiative called Automation of Warehouses because it's a distribution company.
But when they run all of this, they find out, that what creates the most value for their customers was rapid quoting.
So their quoting time was around 18, 19 hours, and customers wanted same-day quotes.
So the CEO had a conversation with the chief operating officer, who was responsible for all warehousing, that we're going to pause this automation initiative.
And instead, he wanted the CEO...
more involved in the coding.
And that had a lot of things that the COO could contribute to.
For example, they needed to upgrade their entire database of inventory to be able to do the codes better, which the COO led instead of the warehousing system.
It's interesting that not only are companies and executives not gathering the data and reading the data, that they're frequently overriding.
In other words, the data is presented to them and they're overriding it based on their experience, their conversations with customers.
I mean, that's more than a blind spot.
Absolutely.
That is such an important point you are making because it goes to self-awareness and humility and appreciation or science in the strategy process.
And if you think about any organization, if you go to the front lines, whether it's like, you know, a welding company, a forging company, or any kind of company, the front line is completely infused with science.
You know, they do little calculations for every step.
As you move higher and higher in the organization, it all becomes salience, hand-waving, and so on and so forth.
So when I work with companies, I always insist that can we do like a day-long session?
on how science can inform strategy.
So I've been like working with this concept called science-driven strategy.
The idea is not that your intuition doesn't matter, that all of that is, you know, useless.
The idea is that it actually can have even more value when it's informed and bound by science.
But it's so interesting, the idea that executives, you know, willfully or otherwise are essentially overriding.
what the algorithms are telling them.
And you have a fantastic line in the article, I love it, which is a statistical model that predicts customer value from survey data is in principle no different from the algorithms Netflix uses.
The difference is that Netflix doesn't convene a committee of experts to override its algorithm every quarter.
I mean, that's a powerful line.
That moved me.
Yeah, yeah, yeah.
And, you know, in some ways, when you have science on your side, You don't need to do quarterly retreats and reviews.
In fact, many of the companies that I've worked with, they have stopped doing those because, you know, they can become over time so confident in the results of science.
Remember the ITI example that we have in the article, right?
When we started working with them, they were at like three, three and a half dollars in terms of stock price.
And, you know, like four years prior to this process, they had gone to five.
And the CEO told, he says, Vakas, we had like a 55% increase in stock price.
We are ready to open up champagne.
And with all of this, whereas, you know, the S&P went up only 35%, their company went up 160% or something like that, right?
That is the power of science, right?
Because the main idea is that it can help you to focus your time, effort, energy, and resources on the few things that create value and excel on it.
Right.
And that's that's a way of thinking rather than trying to do 50 things.
What are the two things I can excel on?
Yeah.
So it's interesting.
Your piece does not suggest that leaders are complacent.
Quite the opposite, that executives change priorities much faster than maybe customers do.
So correct.
How can a leader distinguish between a necessary strategic pivot from just sort of organizational attention deficit?
When you don't know what to do, we just keep throwing things against the wall, hoping something sticks.
But when you know what to do, then you should slow down and become really excellent at it.
And the idea is, how can you train yourselves from running multiple sprints to running a few marathons?
Yeah.
All right.
So if a CEO, if an executive is listening to this and suspects that their company is pursuing too many initiatives, what should happen at the next leadership meeting?
Before they have the next leadership meeting, they should think about identifying what are the one or two things that create the most lift in customer value.
Having a real conversation about what are ways in which you can engage all the employees to do this.
And the leadership meeting should really be about crystallizing that approach.
Talk more about the value.
What does this look like for the companies you studied who did extract true value, who identified the actual value drivers that the data was presenting?
Let me start with benefits to the customer.
So one thing that you see that your customer value scores, and you can measure it a hundred different ways, like customer satisfaction or whatever you want, you see a huge increase in that.
And corresponding to that, you typically see a huge increase in customer retention, sales, margin, cash flow.
In fact, we just finished a paper where we show that customer satisfaction, which is basically a surrogate of customer value, can reliably predict 12 different outcomes for a company, starting all the way from actual repurchase behavior to stock price.
Right.
So that is the first benefit you get.
Like customers can clearly see how you are better than the competition and they become more loyal to you in terms of employees.
So when I start this work with several companies, we measure something from all employees, which is in an average week, the percent of time you spend doing the job that you were actually hired to do.
Take a guess what that percent is for most companies.
35 to 45 percent, meaning all that other time is being done in what they call strategic initiatives, meetings, this, that, and whatnot.
So in ITI, that percent changed.
It was like 25 percent or something like that when we initially started.
And now 70 to 80 percent of people's time is spent doing the job that they were actually hired to do.
And this is a very helpful thing.
And the third is obviously, if you can do both these things, the increase in enterprise value.
I mean, it must get messy at some times.
You know, imagine if your biggest customer demands something that your analysis says has little value more broadly.
What do you do with that?
So one company, for example, their biggest customer is one of the, you know, like the top 10 companies by market cap.
People in that company really did not know that rapid coding was such a big driver of value for a lot of their purchasing managers.
So these people went and educated them.
And just for that company, they created a customized app, which took the coding time instead of five hours to less than five seconds.
So the biggest customers for most companies are also the most savvy customers.
And, you know, there is a thing that customers don't want vendors who are just appeasing them.
Customers want vendors who are actually excelling at the value drivers.
You know, executives think they understand what customers value.
They don't.
They often don't.
So you've already talked a little bit about what somebody would do in a leadership meeting.
But more broadly, how do you get from thinking you know what customers value to really understanding what they value?
The biggest impediment I have seen executives encounter is lack of humility, meaning starting from a place where you can imagine that you could be wrong, right?
I hear this so many times, like, Bakas, you don't understand our business.
You know, I've been doing this for the last 20 years.
You know, we know what our customers want.
If you can just walk them off that ledge and just say, look, There are ways to figure out what value drivers are and prioritize them.
That's all it takes.
It just cast dates.
When I was a teenager, my favorite comedy group was called the Firesign Theater, and they had a whole thing that was basically called Everything You Know Is Wrong.
I feel like we're in that zone here, right?
Yes, because one of the things I did is...
I did a couple of exercises for all these companies.
I had executives rank order the value drivers.
And then I discovered how much value each driver created.
And then I correlated with what executives said versus what actually customers value.
That correlation in the vast majority of the companies, when I say vast, I don't mean 51%, is either zero or negative.
And none of these companies were failing companies, Adi.
They were all reasonable companies doing well.
But once they aligned their strategy to the customers and focused their employees to excel at it, the financial performance improved significantly.
So I love this because it is, you know, at once kind of counterintuitive.
And at the same time, what you're suggesting is very simple.
It's right there in front of our eyes if we choose to take a look.
Yes, absolutely.
Vikas, I want to thank you for being on IdeaCast.
This has been a great conversation.
Thank you so much.
That was Vikas Mithal, marketing professor of the Jones Graduate School of Business at Rice University and co-author of the HBR article, Do You Know What Your Customers Value?
Next week, Allison speaks with Jack Dorsey about building an organization in the intelligence era.
If you found this episode helpful, share it with a colleague and be sure to subscribe and rate IdeaCast in Apple Podcasts, Spotify, or wherever you listen.
If you want to help leaders move the world forward, please consider subscribing to Harvard Business Review.
You'll get access to the HBR mobile app, the weekly exclusive insider newsletter, and unlimited access to HBR online.
Just head to hbr.org slash subscribe.
Thanks to our team, senior producers Mary Du and Rachel Swaby, and senior production editor Kristen Murphy Romano.
And thanks to you for listening to the HBR IdeaCast.
We'll be back with a new episode on Tuesday.
I'm Adi Ignatius.