Here’s a question that sounds easy. How many customers do you have?
We’ve asked it at distributors and dealers where 3 people in the same building give 3 different answers. One says 10. One says 1,000. One says 10,000.
Nobody is lying and nobody is bad at their job. One is counting accounts that bought this quarter. One is counting everything in the ERP. One is counting every contact who has ever been quoted. There is no agreed definition of “customer.” The 3 systems that hold customers were never asked to agree.
Now try to answer a harder question on top of that. Which of those customers are you losing?
I was on a Distribution Strategy Group panel last week called From Order to Insight, with Michael Biwer of Cavallo and Ian Heller hosting. Michael spends his days inside distributors’ order data, and he said something early on that I keep coming back to.
The data has been there all along. What’s been missing is the intention and the technology to read it.
The ERP is the operating system, not the app
Michael has a comparison I’d steal if I thought I could get away with it. The ERP is to your business what the operating system is to your phone.
You need the operating system. Nobody argues about that. But nobody chooses a phone because of it either, and nothing you do all day is the operating system. It’s the apps.
The consequence is uncomfortable if you’ve spent real money on an ERP: it isn’t where your advantage comes from. Your competitor can buy the same one. These systems are built to serve a flower shop, a manufacturer, a nonprofit, and you. By design, they aren’t built to give any one of you an edge.
Michael’s way of putting it is the one I keep repeating: the ERP records what happened. It is built to process and record transactions, not to judge whether each one was a good one.
Here is where I’d push it further. It doesn’t record why it happened, or how. Somebody agreed that price. Somebody chose that substitution. Somebody decided this customer could have net 45. The order is in the system. The reasoning never was — it was in a conversation, or on a sticky note, or in one person’s head.
Your ERP records the order. Your CRM records the relationship. Neither one reads the economics underneath: the account whose mix quietly got worse, the supplier that raised its price 4 times this year while your sell price never moved, the rep who discounts harder than the rest with no floor set.
Ian put it as well as anyone has. You can own an acre with gold under it. If you don’t know and never look, the gold doesn’t matter.
Looking after the fact, versus in time to do something
Most distributors and dealers who look at this at all look at outcomes. The month closes, the report runs, and the margin is what it is.
Michael’s framing of the alternative is the sharpest thing I heard all hour. The difference isn’t visibility after the fact. It’s visibility with the chance to intervene.
Finding out today about something that happened yesterday is useful — you can still act. Finding out while the quote is still open is better, because then the thing never happens. A price that was never wrong doesn’t need a credit memo, a conversation with the rep, or an apology to the customer.
And this is where the churn argument gets practical. Michael’s point was that being good at predicting churn isn’t enough if you predict it late. A model that tells you accurately that a customer is gone is a model that told you after the useful window closed. What you want is the earliest signal you can act on.
Michael put the everyday version of it to us like this: a lot of companies wonder who their sales team should call today. His answer was that it should be very clear, on a daily basis, from the data they already have. For most it’s answered from memory instead — and memory calls the accounts that are pleasant to call.
Don’t wait for clean data
Here’s where I expect most people reading this to bail out. I hear the same sentence constantly. Our data is a mess, so none of this applies to us yet.
Ian said the next distributor who tells him they have good data will be the first one ever. I believe him. I’ve never met that distributor either.
Michael had the best answer to this that I’ve heard. He described a boat with a leak, filling up with water. Someone is standing in it, refusing to act until they know the precise dimensions of the hole.
Precision is not the same thing as progress. You almost certainly have enough good data today to make a better decision than you’re making now — not every decision, but the expensive ones. Waiting for clean data is how a project stays a project for 3 years.
There’s a related trap on the other side of it, which is trying to fix everything at once. Michael’s advice was to prioritize the problems with the greatest economic consequence instead of optimizing everything simultaneously, and he was blunt about why: that’s why most initiatives fail. The 2 things he said they fail for are a lack of clarity and conviction about what’s being solved, and no real change management or leadership behind it.
He closed the hour on a phrase he said he’d picked up somewhere and couldn’t source: are you in love with the problem? Because the odds of solving it go up a lot if you are. I’ve been repeating it all week.
Staple yourself to an order
If you want one place to start, I’m passing on something I got from Michael on the panel. It isn’t his, and he was clear about that. It’s a Harvard Business Review classic called “Staple Yourself to an Order.” He says a former chairman quoted it at him once, when he was describing his own business. Ian looked it up live and put the link in the chat. The idea is simple: follow one order all the way through your own company and watch what happens to it.
Michael’s addition is the good part, and it’s the bit I’d actually do. Don’t just staple yourself to the order. Carry a stopwatch, a calculator, and a camera.
The stopwatch tells you where the time goes. The calculator tells you what each step costs. The camera tells you what the thing actually looks like — the screens, the clicks, the printout somebody walks down the hall.
And then one more line of Michael’s that I’d put above all of it: don’t start by asking what data you have. Start by asking what economic problem you need to solve. Then go and find the data that answers it.
That order matters more than it sounds. Starting from the data gets you a dashboard. Starting from the problem gets you a decision.
What “one view” actually takes
I’ll say the unglamorous part, since it’s the part I work on.
Getting to one view of the business is not a reporting project. It’s a definitional one. Before you can count customers across an ERP, an ecommerce platform, a CRM, and a warehouse system, something has to decide 3 things. What a customer is. Which system wins when 2 disagree. How confident you are in each.
That’s the work. It isn’t exciting and it doesn’t demo well. But everything else sits on top of it — the churn signal, the margin control, the AI that’s supposed to be smart about your business — and none of it can be better than the layer underneath.
Where I’d start
- Agree on one definition of a customer. In a room, out loud, with the people who disagree. You’ll learn more in that hour than from the next report.
- Pick the one economic problem with the biggest consequence. Michael’s, and he’d tell you it’s the one most people skip. Not the longest list. The most expensive line.
- Staple yourself to an order. Michael’s, by way of HBR. Stopwatch, calculator, camera.
- Ask what you’d do differently if you knew sooner. If the answer is nothing, it was never the problem. If the answer is obvious, you’ve found where to put a control.
The panel is available on demand. Michael’s half of it is where the order-data argument sits.
If you’ve done the exercise above and the stopwatch told you something you didn’t want to know, I’d like to hear it. Those conversations are the most useful ones I have.
— Shawn

