There’s an interesting paradox in e-commerce: We have more data than ever, and nearly infinite ways to analyze, but brands have never felt less certain about where to spend the next marketing dollar.

Every dashboard, platform and AI tool promises to be a definitive source of truth: This channel drove that revenue, this campaign has a 4.2x ROAS, this creator drove $X in revenue, etc.

But as you’ve probably seen, that’s just not reality.

The dashboards never agree, the numbers don’t add up, and the story changes depending on who built the report.

But the answer isn’t to throw it all out and just YOLO, or to chase the impossible dream of flawless attribution.

The solution is nuanced, but simple:

Build a system for measuring that’s honest about what we don’t know but still robust enough to make smart bets on profitable growth - and we’ll go over exactly how to do that.

Signs of Attribution Chaos

But first, let’s ask the bigger question: What are the signs of trouble? If any of this looks familiar, it may be time to rethink your approach to attribution.

Channel whiplash

A brand sees Meta ROAS fall off a cliff in‑platform. They panic, slash spend, and get excited about cutting “inefficient spend.” And in the immediate term, the dashboard looks better!

But they quickly realize it was an illusion:

First, overall revenue drops. Then organic and branded search, email metrics, etc. Essentially, all the core KPIs start to slide, and panic sets in.

The attribution dashboard said Meta (or Google, TikTok, etc) wasn’t driving value, but it’s pretty clear that doesn’t tell the whole story.

Those Meta ads may not show a conversion, but millions of people seeing your brand, product, etc every month creates a lift in everything that’s hard to precisely track, but very real.

Attribution tug‑of‑war & paralysis

The digital marketing lead shows a dashboard where paid social is printing money. Great news! But then the brand lead counters with a dashboard showing direct and branded search doing the lifting… who’s right?

And then you look at analytics and see a huge chunk of revenue attributed to “direct” or “unassigned.” Is that REALLY from direct, or is it mis-attributed?

The truth is that there is no objectively right answer. Everyone has their own (arguably correct) interpretation.

So with conflicting stories in the data, brands often just default back to hunches.

The result: Short‑termism and channel addiction

This all adds up to one of the biggest things holding brands back from growing:

Because last‑click and in‑platform metrics favor easy‑to‑track conversions, teams drift into over‑investing in retargeting and branded search, and under‑investing in content, creators, and brand campaigns.

The cycle looks like this:

Cut upper‑funnel because it doesn’t look good in the dashboards, watch overall demand soften, then spend more on bottom‑funnel tactics to squeeze what’s left.

Inevitably CAC rises, email campaigns stop performing, and profitability tanks. Which makes the problem even worse, because the brand can’t afford to invest in the experiments that will find new ways to grow!

We’ve probably all seen it… but the question is, how do we fix it?

Three ways out

The answer isn’t in hunting for the tool that solves marketing attribution - that’s a pipe dream. You fix it by changing how you think about measurement.

1. Use a measurement portfolio, not a single source of truth

Instead of chasing the impossible dream of the one “perfect” dashboard, use a portfolio of imperfect tools:

Platform data as directional, not gospel. If Meta says your CPA is spiking, that’s a signal to test and investigate - not an automatic kill switch.

Blended business metrics as anchors: new and returning customer orders, MER, blended CAC, contribution margin, payback period, LTV/CAC by cohort. These are hard to game and reflect reality at the business level and aligned with the P&L.

Simple mix and incrementality tests: on/off or heavy‑up tests by channel, basic geo holdout tests, brand vs. direct‑response splits at constant spend.

None of these are perfect alone. But when multiple lenses show you a similar story, it’s probably valid.

2. Move from conversion metrics to decision metrics

Most fights happen because people try to answer business questions with campaign‑level metrics, eg greenlighting a new SKU because they saw some juicy ROAS last week.

Instead, zoom out and start with the real decisions:

• How much can we profitably spend this month and quarter?
• What’s happening to CAC, MER, and payback as we change channel mix?
• Is new customer acquisition getting better or worse over time?

Then use channel‑level metrics to help answer those questions. Sure, ROAS is one metric. But also look at the share of new vs. returning customers, impact on branded search and direct traffic, blended CAC and any other data you need for an informed decision.

For brand and creator work, look at search volume, site traffic, list growth, and the baseline of your performance campaigns.

Then you get past the dogfights over “Should this ad get credit for the sale?” and get to “Did this investment make the business healthier?”

3. Make experimentation part of planning

Almost every company says they believe in testing new creative, new audiences, etc. But in practice, experiments are usually the first thing cut.

It’s understandable: they feel like a “nice to have” vs a “must have.” But if you’re not testing, you won’t find new growth areas and it’s only a matter of time until you plateau and CACs climb higher every month.

So instead of cutting experiments, try this:

• Set aside a fixed percentage of marketing spend for structured tests and protect it.
• Define each test up front: what you’re testing, how you’ll judge success, and the time window.
• Keep tests simple and repeatable: turn a channel on/off, creative concept vs. control, brand‑heavy vs. DR‑heavy at constant spend.

You’ll never have a “perfect” answer, but that’s ok. The point is to continuously learn, so every quarter you get more confident about where the next dollar should go.

The bottom line

If you wait for a clean, universally trusted attribution tool before making bold decisions, you’ll be waiting forever. The platforms will keep changing; the data will never be perfect.

The winning brands aren’t the ones with the fanciest dashboards. They’re the ones who accept uncertainty, use multiple imperfect views of reality, anchor on business outcomes, and run disciplined experiments to keep getting smarter over time.