Data

The foundation everyone wants to skip and no one can afford to

Every revenue problem I have ever been handed came with a theory attached. Pipeline is down because the SDRs aren’t prospecting enough. Conversion dropped because marketing is sending junk leads. The forecast is wrong because sales is sandbagging. The theories are always about people, and they are almost always wrong, because the real problem is usually sitting one layer underneath, in the data, quiet and unglamorous and expensive.

This is the first layer of the Revenue Teardown, and it is first for a reason. Everything else in your revenue system is built on top of your data. Your scoring model reads from it. Your routing rules depend on it. Your attribution is calculated from it. Your forecast is an interpretation of it. If the data underneath all of that is wrong, then everything above it is confidently, precisely wrong, which is worse than being obviously wrong, because at least obvious wrong makes people cautious.

So before I look at anything else, I ask one question: can you trust what your systems are telling you?

Most of the time, the honest answer is no. And most of the time, nobody wants to hear it.

What “bad data” actually looks like

When people hear “data quality” they picture a few duplicate contacts and a couple of blank fields. If only. Here is what it looks like in a real, mature instance that has been running for years.

You have the same account existing as three different records, because someone typed “IBM,” someone else typed “I.B.M.,” and a third person let an integration create “International Business Machines Corp.” Your rep is working one of them. Your marketing automation is scoring another. Your renewal team is looking at the third. None of them know the other two exist.

You have an “account owner” field that is populated for sixty percent of your accounts and blank or wrong for the rest, so your routing rules quietly drop leads into a void where no human ever sees them. You will not notice this, because nobody gets an alert that says “a lead you should have worked was never assigned to anyone.”

You have fields that mean different things depending on who filled them in. “Region” means sales territory to one team and physical office location to another. “Customer” means closed-won to finance and means “has ever been in an opportunity” to marketing. Every report built on those fields is subtly, invisibly lying.

And you have data that was true once and is not anymore. Contacts who left their companies two years ago. Phone numbers that ring nowhere. Job titles from a promotion cycle ago. Data doesn’t just start wrong; it rots.

Why this is the layer nobody wants to fund

Here is the uncomfortable thing about data work: it is invisible when it goes right and invisible when it goes wrong, until it is a catastrophe. A clean CRM does not generate a press release. Nobody gets promoted for deduplicating accounts. There is no dashboard that lights up green and says “your foundation is solid.”

So when a leader has budget to spend, data cleanup loses every time. It loses to the new intent-data platform that has a slick demo. It loses to the extra headcount that feels like visible progress. It loses to the rebrand, the new campaign, the shiny thing. Data work is eating your vegetables, and every organization I have ever seen would rather order dessert.

I understand the impulse. I have felt it myself. But I have also watched what happens downstream, and it is always the same. The expensive new tool gets bolted onto the messy foundation, inherits every problem in the data, produces garbage, and six months later everyone is confused about why the thing they bought didn’t work. It didn’t work because you asked a rocket engine to run on dirty fuel.

A story about a nine-times return

Early in my time at Life Line Screening, the marketing numbers looked broken. Spend was going up, conversion was not following, and the working theory — of course — was that the campaigns were bad. Fix the creative, fix the targeting, fix the people running it.

I did not touch the campaigns. I went underneath them, into the Marketo instance and the data feeding it, and what I found was a foundation that had been built in a hurry and never maintained. Duplicate records everywhere. Fields that had been repurposed three times. Sync rules to the CRM that dropped or scrambled data on the way through. The campaigns weren’t failing because they were bad. They were failing because they were built on sand.

So we rebuilt the enterprise Marketo instance from a clean foundation. Not the creative, not the targeting, the foundation. Deduplicated, redefined fields with actual agreed meanings, fixed the sync so data arrived intact. Unglamorous, months of work, nothing you could put in a highlight reel.

The result was a nine-times return on marketing investment. Not because we suddenly got clever with campaigns, but because the campaigns we already had could finally be measured, targeted, and trusted. The fuel was clean, so the engine ran.

That is the pattern. When you fix Layer 1, the layers above it often start working on their own, because they were never actually broken. They were just starving.

How to check your own foundation

You do not need a six-month audit to know whether you have a data problem. You need to ask a few blunt questions and be honest about the answers.

Pick your ten most important accounts. Can you, in under a minute each, find the single authoritative record, see who owns it, and trust every field on it? If you find duplicates, blank owners, or fields you don’t believe, that is your answer.

Ask your best rep and your marketing lead to define “qualified lead” separately, in writing, without talking to each other. If the definitions don’t match, your data is encoding two different realities.

Look at your last forecast miss. Trace it back. More often than not, the miss started as a data problem — a stage that meant different things to different people, a set of records that were double-counted, a pipeline that included things that were never real.

Fix this before you touch anything else

If Layer 1 is broken, you stop here. You do not tune scoring, you do not buy intent data, you do not rebuild attribution, because all of those things read from the foundation you have not fixed yet. Doing higher-layer work on a broken data layer is how companies spend a year and a lot of money and end up exactly where they started, just poorer and more confused.

It is the least exciting recommendation I give, and it is the one that pays off the most. Clean the data. Own the records. Agree on what the fields mean. Then, and only then, build on top of it.

The foundation is boring. The foundation is everything.