The Secret Sauce
Back when I sold equity research, I opened every sales pitch by asking people to think about the idea of “signal vs. noise.” My clients — distracted, stressed-out hedge funders —always seemed to pay a little more attention during that part of the meeting.
“There’s a lot of information out there on the companies you cover,” I would say. “How are you cutting through all of that info to figure out what’s really going on inside the business?”
My pitch wasn’t bullshit. Our company was built on deriving signal from noise. Our research analysts gathered, tracked, and analyzed proprietary datasets for every company we covered. Then we correlated that data to metrics that moved stocks — revenue, pricing, transaction count, and earnings-per-share. We analyzed credit-card data for companies like Netflix, pricing data for airlines and cruise lines, and online home listing data for the homebuilder stocks. Our product’s magic wasn’t any one particular data point. Our approach worked because we selected just a few important sources of information for each company and then reviewed them daily — again and again — to help us objectively derive “what was really going on.”
This simple, two-step process — select and review — is what kept our research team objective. It helped us figure out what others might be missing. The Wall Street analysts we competed against touted their years of experience and their chumminess with management teams. Their legacy and their relationships were their secret sauce. Our secret sauce was the data. More specifically, our secret sauce was our disciplined focus on a handful of data points that mattered for each company and our willingness to stare at that data while asking a single question:
“So what?”
How to Hide
I left my sales gig after a while to get my MBA. In my last decade as a strategy consultant and PE advisor, I’ve noticed something unfortunate. Most businesspeople don’t use data to make decisions. They use data as a shield. When these people do share their numbers, they generally have some sort of ulterior motive. They’re not actually trying to make things better. They’re trying to manage their status, appease their superiors, and project a calming signal that everything is under control. This makes me sad. It also makes my job more difficult. I like helping businesses fix things. And it is almost impossible to fix a secret.
This plague of unfulfilled analytical potential is a shame. We’ve come such a long way. At least our technology has. CRMs, business intelligence platforms, and the giant informational butterfly net we call the internet has made it so much easier to gather and make sense of the data that explains how a business works and how to improve it. It’s easier than ever for anyone inside a business to understand how things are going, what’s getting in the way, and what we might be able to do about it. There’s no excuse anymore to not know what’s going on.
Despite all this analytical potential, too many managers treat their data like defense attorneys treat evidence: It’s only there to get them out of a jam.
Some all-too-common examples:
The CRO wants to get the CEO off their back, so they cherry-pick a few metrics that show their sales projections are likely to pan out. See? We’ve got more than 5x pipeline coverage. We’re fine.
The Head of Marketing doesn’t have time to analyze if the new campaign is working, so they snapshot the highest-level traffic count they can find. See? Look how many people visited our website last month! Things are going great.
The Engineering leader wants to show that your product works, so they pull their NPS number for the first time in a few months. See? It’s still positive. All good over here.
Most managers only look at their data when they absolutely have to. And too many use numbers to protect themselves — not to diagnose problems, not to make decisions, and certainly not to improve their business. They know the information is there. They just aren’t using it.
These people are being data-aware — not data-driven.
Being data-aware isn’t enough. Data is only as good as the dialogue it creates.
Why U No Like Data?
Most managers aren’t dishonest. They’re just afraid. They don’t know what data to look at, how to analyze it, how to pull out the insights that matter, or how to talk about them in front of senior executives. No one ever helped them achieve data literacy. No one ever taught them how.
Steve Levitt, author of the best-selling book Freakonomics, has some thoughts on this. In an episode of his podcast, titled “Americas Math Curriculum Doesn’t Add Up”, Steve examines the types of mathematical chops we learn in US schools and how they contribute to a widespread lack of data literacy.
It’s a great podcast. Partly because Steve’s point is so simple: We’re teaching our kids the wrong analytical stuff. Steve wants our schools to replace abstract topics like trigonometry and geometry with more pragmatic, data-focused coursework. You know, the kind of math you can actually use. I think he’s spot-on. If we spent less time on the Pythagorean theorem and more time learning to crunch real-world numbers in a simple spreadsheet, maybe the average U.S. small business (and the U.S. consumer) would be in a better place right now.
Unfortunately, Steve’s ideas face a long, uphill road before they make their way into our classrooms. Tectonic plates move faster than curriculum reform. I’m not holding my breath waiting for high schools to start teaching pivot tables. And if you’re in the working world already, it’s time to wake up. No one is going to close this gap for you. So as you’re reading this and quietly admitting to yourself, “Yeah, I probably need to learn how to be more analytical,” then I’d like to help.
Here’s how to get started.
[Step 0] Please. Learn. Excel.
You don’t need to be an analytical whiz to be data-driven, but you do need to be comfortable cleaning, manipulating, and formatting basic information about your business. Yes, in a spreadsheet.
…So, can you do that? This is a gut-check moment. If you’re not able to take an unformatted flat file and build a simple spreadsheet that answers a question (using formulas for basic arithmetic, data sorting, and pivot tables) you will always feel behind.
So start there. Close that gap. There are plenty of courses that can take your excel skills from zero to useful in just a week or two. If you’re in sales, ClozeLoop’s Spreadsheets for Salespeople is a great self-study option. If you work elsewhere, ExcelJet is a fantastic place to start. There are plenty more out there. Take it upon yourself to build the foundation. Without it, you’re going to struggle. (Note: If you ask nicely, your company will probably pay for an online course. Most people never ask.)
[Step 1] Answer Three Questions.
Being data-driven is about selection. It’s about picking a few important questions for your area of the business, and then selecting easy-to-get sources of information that (i) objectively answer those questions and (ii) help you make better decisions. That’s it.
Asking the questions kickstarts the whole process. I like to start with three basic types.
Results Questions: “How did we do?”
Predictive Questions: “How will we do?”
Quality Questions: “How well are we doing it?”
Be careful. There’s a speed trap here. You could answer all three of those questions right now, subjectively, using your feelings, your intuition, and what you’ve seen recently. An example: Maybe you’re a sales manager. Your team just booked a few big deals last week, you have a verbal commitment from another customer set to close next week, and you just sat through a demo of your product that seemed to hit all the right notes. We’re doing great! Seems like we’ll continue to keep doing great! My team is killing it!
…But are they, though? What about that sales rep that hasn’t closed anything in 6 months? What about your measly 1.8x pipeline coverage? What about that discovery call that’s going off the rails, out of your field of view, at this very moment? As I’ve written before, the tough part about becoming a manager or executive is this: No one tells you more responsibility comes with less control. There is so much you can’t observe for yourself, and so many opportunities to make things better for your team that you will miss if you rely on your single, biased, limited vantage point. Peter Drucker nailed it way back when. “If you can’t measure it, you can’t manage it.”
That’s why answering these questions objectively (using numbers!) is so important. It’s a form of managerial humility. The data-driven manager isn’t trying to dazzle people with dashboards KPIs or appease their more quantitatively-inclined peers. They’re looking for a way to answer these questions as truthfully as possible — so they don’t miss an opportunity to help their team close a gap, beat a target, or raise the bar on how they do their work.
Here’s a great place to start — using the table below, pick just one metric that answers each of the three questions. Then set a target. (Far too many managers out there report out numbers without any point of reference — without any way to tell if they’re winning or losing. Is 51 MQLs bad or good? That question answers itself when you set a target in advance.)
An important note: The metric + target that you choose should be easy to gather. Reports with difficult-to-gather metrics are like signing up for a fancy new gym 45 minutes across town: After a while, you’re just going to stop showing up.
Here’s the template filled out with some example metrics for a sales manager.
A few notes on what’s going on here:
The “Results Question” row prevents you from using “we did the best we could” as an excuse. The results are right there, in black and white. Reminding people of your goal and how you did against it might seem basic, but it’s not. People need reminding.
The “Predictive Question” row houses your leading indicator. What can you track that tells you if you’re on track? Pipeline predicts bookings. NPS predicts churn. Employee engagement scores predict attrition. Pick a number that helps you forecast what you can expect in the future. Put it here.
The third row — the “Quality Question” — is usually the toughest. My coaching here might sound antithetical to the rest of this essay, but here it goes: Emotions are a powerful clue. Use them. What do you have strong opinions about in your work? What do you believe in doing “the right way?” What makes you cringe when you see someone doing it wrong? Those are all great signs of something that deserves a measure of quality that you can be tracking. The best quality metrics are often a measure of preparation (“Did we get ourselves sufficiently ready?”) or conversion (“Did we move forward or get stuck?”).
Don’t overcomplicate this. Fill out the table, pick three metrics, set some targets, and start tracking them every week.
[Step 2] Create + Share a Weekly Report
Your calendar drives your work. The most data-driven C-level people I know block off portions of their calendar to do nothing but review and analyze their numbers. They force themselves to create some kind of summary or report that highlights the “so what’s” for that week — what stands out, what’s changing, and what gaps need closing. The really good ones share that report and their takeaways with their leadership team (and often, with their investors). Andy Grove, CEO of Intel, would have applauded this practice — even if nobody reads the report. After all, as Andy put it…
“Reports are more a medium of self-discipline than a way to communicate information.” Andy wrote in High Output Management. “Writing the report is important; reading it often is not.”
Too many managers are quantitatively sleep-walking. They prepare and share their board slides, their sales kickoff materials, or their budget presentation, all laden with numbers. Then they forget all about those numbers and go back to running things by gut feel. To keep yourself objective and data-driven, you need a forcing function. My advice is to create the expectation in others that they’ll be seeing data from you every week. Every week, you should share a consistent update on your key metrics and how they’re moving. A simple table is fine. The format is not important. The consistent visibility and the discipline of synthesizing and sharing the data is.
You stand a much better chance of improving that which you understand. Want to prove to me that you understand your part of the business? That I can trust you to make it better? That you deserve your seat on the leadership team? Show me a weekly report and share your thinking on why the numbers are moving the way that they are.
Then I’ll believe you. Not before.
[Step 3] Take Stances and Make Moves
All this talk about being data-driven doesn’t actually get you anywhere unless you use what you learn and take action. After you start tracking your key metrics every week, you’re going to notice things. You’ll notice when numbers move up and when they move down. You’ll notice when you expected your numbers to move up or down — and then they didn’t. In your weekly report, you should include analysis (“What’s going on?”) and a stance (“So what? What should we do about it?”) for any meaningful move, trend, or insight you notice in your numbers. Be selective. Time-box it for people. “If you only have 5 minutes to read this, here’s what you should take away.”
Taking a stance can be tricky. I think there are basically three types of stances:
Do nothing — “Something is happening in the numbers, but it either doesn’t matter or it’s out of our control. No action is needed.”
Wait-and-see — “It may be too early or too noisy to tell if a trend in the numbers is meaningful. For now, we should continue to watch carefully.”
Take action now — “There’s a gap to close or opportunity we can capitalize on here — and this is how we do it.”
An explicit stance is a two-way insurance policy. It’s a check on yourself. It ensures you’ve thought through what you’re seeing in the numbers before making a recommendation. But it’s also a signal for your team that communicates openness to input. “Unless other chime in,” it says, “this is what I’m going to do next.” Being explicit about (i) what you see in the numbers, (ii) the “so what’s” and how you’re interpreting them, and (iii) what you’re planning to do about it — that’s the simplest recipe for data-driven management there is.
Tell the Truth
When I was young (like, learning to talk young) my parents would play this game with me. They would repeat a few short phrases until I could finish the last bit of the phrase myself.
Over the holidays this year, we brought out a couple old family videos and watched them together, sprawled out on the living room floor in front of the fireplace. On one of those videos was a short clip of me and my Dad, playing our game.
Dad: “Paul, you ready? What do we say? Always tell the…”
Tiny Paul: “…Truth!”
That was good advice. Not just on how to be a good person, but on how to use information in the right way. Honesty really is the best policy: In life and in data. The best people in any profession search for and share the truth. They drive towards a rigorous, objective understanding of what’s really going on. Then they talk about and fix the stuff that’s broken — the stuff you can’t help but notice after you really dig in and use the numbers.
So don’t stop at being data-aware. Be data-driven.
Don’t just share information. Tell the truth.
Pick a few metrics (in advance) that tell you objectively what’s going on. Schedule time to digest the numbers each week. Build a weekly report. Share your report and the accompanying “so what’s” with your team. Take some stances.
Then ask your team to help you decide what to fix.
Then go fix it.
Would love a 2025 update on this!