I worked with a company that was sure email was driving their sales.
The dashboard backed it up. Source: email. Last touch before purchase: email. The board kept doubling down. Bigger lists. More frequent sends. Same template, slightly different subject lines.
The thing is - the emails were spam. Mass blasts, no personalization, the kind of generic value props you’ve read a hundred times. Anyone who already wanted to buy, for completely unrelated reasons, would click one of those emails on the way to a decision they’d already made.
The dashboard logged that click as the cause.
I keep coming back to this case because the longer I sit with it, the less I understand where the certainty came from. The board kept saying “but it’s hard data.” Hard in what sense? Hard because there was a lot of it? Hard isn’t the same as true.
Eighteen months in, that company still couldn’t tell you why their customers actually bought. They had a much bigger email list to show for it though.
Every founder knows that correlation isn’t causation. Almost nobody applies it when they’re looking at their own dashboard.
Here’s what it looks like in practice. Your prospect goes through a buying process that takes three to nine months. During those months they touch your brand maybe twenty, thirty times. They see a LinkedIn post. They overhear someone mention you on a podcast. They lurk on your site without filling out a form. They ask a peer in a private Slack channel. They read a competitor’s comparison page that mentions you. They subscribe to your newsletter and never open it. Then one day they click an email and book a demo.
Your attribution catches one of those touches. Maybe two. The other twenty-eight are invisible.
So when you look at your dashboard, you’re not seeing the touches that did the work. You’re seeing the touches that happened to be measurable. And then you make budget decisions based on that, and tell yourself you’re being data-driven.
This is the part I find genuinely hard to communicate to founders sometimes. The dashboard isn’t broken. The dashboard is doing exactly what it was built to do - show you what it can see. The problem is what it can’t see is usually where the actual demand was being built.
Email click-through rates. Paid ad CTR. MQL-to-SQL conversion. Form fills. Webinar attendance. All easy to measure.
How a podcast appearance shifted perception in your target market. Whether your CEO’s interview pulled three accounts into the buying cycle six months later. Whether the LinkedIn post you wrote at 11pm last Tuesday is the reason someone trusts you. Almost impossible to measure cleanly.
Guess which list gets investment. Guess which list gets cut because “we can’t prove ROI.”
The list that gets cut is usually where most of the demand at seed to Series A is actually being built. But because it doesn’t show up clean in attribution, it gets defunded. Then a quarter later pipeline starts thinning, nobody can explain why, and the response is - surprise - more email.
I’ve watched this loop happen at four different companies. It’s almost identical every time.
Here’s the part that’s harder to admit.
Even if attribution worked perfectly - which it doesn’t - you wouldn’t have enough data to draw conclusions at this stage.
B2B SaaS at seed to Series A means small numbers all the way down. TAM of three thousand. ICP of three hundred. Maybe four thousand monthly visits to the site. Twelve demos a month. Four closed-won deals a quarter.
I tried running a clean A/B test on a landing page once at numbers like these. Waited a few weeks. Got an “18% lift.” Felt good. Rolled it out. A month later the effect was gone - turns out it was noise the whole time.
I didn’t need a better tool. I had too few people on the page for any test to mean anything.
It took me a while to actually accept this. The instinct as a marketer is always “let me set this up properly and the data will tell me.” At seed to Series A, the data won’t tell you. There isn’t enough of it. What you get is just enough numbers to be confidently wrong.
This is the trap. Founders demand data-driven decisions because data-driven sounds responsible and grown-up. But there’s no data to drive with. There’s just enough to mistake noise for signal, and double down on whatever the dashboard happened to highlight last quarter.
The board in the case I opened with kept saying “but the data shows email works.” Email got clicked. That’s a different sentence than “email caused the sale.” Nobody at the company was asking the second question. They were asking why dashboards lit up. Those have completely different answers.
The tools you’re paying for were already weak. Now they’re crippled.
Browsers block third-party cookies. iOS strips UTMs. GDPR pop-ups get rejected by most visitors. Server-side tracking helps a bit but doesn’t recover what’s lost. And dark social - buyers researching in Slack channels, in DMs, on podcasts they listen to in the car, in conversations you’ll never see - leaves no trace at all.
Most B2B buying activity now happens before the prospect ever talks to your sales team. None of it is in your CRM. None of it is in Google Analytics. Most of it is happening in places your tracking pixel will never reach.
There are thousands of attribution tools on the market trying to solve this. Multi-touch attribution, marketing mix modeling, identity resolution, AI-this and AI-that. They all share one feature - they don’t show you the truth. They show you a model that approximates the truth, with error bars wide enough to drive a strategy through.
If you’re spending money on better attribution tools, you might be solving the wrong problem. The problem isn’t tooling. It’s that you’re trying to extract confident conclusions from data that fundamentally can’t support them.
This is where most articles like this collapse into “trust your gut” or “do brand marketing.” Neither is operational.
The actual answer is harder and more boring. You build a process for talking to people. The same way good product teams build a process for customer interviews, B2B marketing at this stage needs structured ongoing dialogue with three groups: your sales team, your customers, and the prospects who didn’t buy.
Not occasionally. Continuously. Built into the operating cadence.
Most companies have a sales-marketing meeting that’s actually a status update with grievances. Marketing reports MQLs. Sales says the leads are bad. Marketing says sales doesn’t follow up. Nothing useful happens.
Replace it with something tight. Twenty minutes. Three questions:
What patterns showed up in discovery calls this week? Marketing needs to hear the language buyers actually used. Not in a quarterly readout. This week.
What’s missing? What questions does sales hear that marketing isn’t answering anywhere - on the site, in content, in collateral? Every gap is a content brief.
Where did good prospects come from this week, in their own words? When sales asks “how did you hear about us” and the prospect says “a friend mentioned you” or “I saw your CEO on a podcast” or “I’ve been reading your VP of Sales on LinkedIn for months” - that goes into a shared doc. That’s your real attribution. Not the dashboard.
This isn’t a meeting that needs slides. It needs honesty and a shared note-taking habit.
Three groups, on a rotating schedule:
Current customers. Why did you pick us? Walk me through the two months before you reached out. What were you reading, watching, asking peers? What almost made you go with someone else?
Closed-lost prospects. Same questions. Plus - what would have changed your mind?
Churned customers. What did we promise that we didn’t deliver? What did we deliver that we underpriced? Where did you go instead?
Each interview is forty-five minutes. Recorded. Transcribed. Tagged for patterns. After ten interviews you’ll know more about why people buy from you than your dashboard has shown you in three years.
I’ve tried doing this in different ways. Heavy structure with formal questionnaires. Loose chats over coffee. Recorded with full transcripts. Just notes. What worked best for me was something in between - light structure, lots of room for the “actually, I want to add one more thing” moments. Those are usually where the gold is.
The hardest discipline at this stage is refusing to draw conclusions from inadequate data.
If you have one channel that “delivered” three deals and four channels that “delivered” zero - you don’t know that the first one works. You don’t know that the others don’t. You have a sample size that justifies no conclusion.
Better to say “we don’t know yet, let’s keep all five running and listen to what customers say” than to cut four channels based on noise. Cutting channels based on bad data is the most common way I’ve seen B2B SaaS marketing teams accidentally kill the things that were actually building demand.
“I don’t know yet” is a complete and respectable answer at this stage. “The dashboard says X” - often isn’t.
I run MOAT to help B2B SaaS founders build account-based marketing capability inside their company instead of renting it from an agency. The pattern in this article shows up in almost every conversation I have with founders.
The fix isn’t a better tool or a more rigorous attribution model. It’s a marketer inside your company who’s close enough to your sales team and your customers that they don’t need to rely on the dashboard for direction. Someone who walks into the weekly with the actual language buyers used in last week’s discovery calls. Someone who’s done seven customer interviews this quarter and can tell you, in specific quotes, why your last three deals closed. Someone who knows when the data is too thin to act on, and has the judgment to say so out loud.
Agencies can’t do this. They’re not in your business often enough.
Fractional consultants can’t do this either. They’re not in your business deeply enough.
A senior ABM hire can - eventually. But at seed to Series A you usually can’t afford one yet.
What you can do is take the mid-level marketer you already have, or hire the right one, and build them up with external pattern recognition until they have both context and judgment. That’s what I do at MOAT. Not for them. With them.
If your dashboard has been telling you a story for six months and your pipeline doesn’t match - that’s the signal. The story is wrong. Not because the data is bad. Because data alone was never going to tell you the story you need.
If you want to think through your situation, I run a free 45-minute diagnostic call for founders who fit the profile. We’ll look at how your team is currently reading attribution data, identify a couple of channels you might be underinvesting in because they don’t show up clean in your dashboard, and talk about what the build-out actually looks like. No pitch, no deck. You leave with useful direction either way.
About the author
Daniel Tupczyński runs MOAT Marketing, a small consultancy that helps B2B SaaS founders build in-house ABM capability instead of renting it from an agency. He also leads marketing at a B2B SaaS company.
