Marketing measurement for DTC brands

Are your platforms taking too much credit?

I help growth teams test what their marketing actually changed—then build the systems that keep the learning going.

Brinker Analytics is my independent measurement practice for DTC brands spending $1M+ a year on ads. I use MMM, geo tests and attribution to help you decide where the next dollar goes.

Start a conversation

70+: incrementality tests
$123M: personally launched and run in digital campaigns
~$60M/year: ad spend directed, 2023–2026
10+ yrs: in DTC marketing analytics

The problem with certainty

How do you know which marketing is actually working?

It depends on the decision you're trying to make. Each platform claims its own sales, so the reports can add up to more orders than you took. You find out what's working by putting the evidence side by side. Attribution shows the path you can see. An MMM covers the whole portfolio. Tests break ties. When the reads agree, you can act. When they disagree, run a test before you move the budget.

Should you spend up, spend down, pause a channel, or change strategy? Facebook says it drove 500 sales. Google says 400. You only had 600 total orders. The reports disagree because they are all looking at the same messy reality from different angles.

The answer isn't to pick a winner. It's to put the evidence in conversation.

The Measurement Council

Triangulation, not model wars.

No single source of truth. A disciplined cross-examination of every credible read.

Models

Portfolio-level perspective

MMMs put spend, seasonality, promotions, and demand into one view. They can hear portfolio signals that click data cannot.

Experiments

Causal calibration

Geo holdouts, matched markets, and synthetic controls test what changes when spend changes. They keep the models honest.

Behavioral data

The journey, in detail

MTA and first-party identity data reveal observable paths to purchase and surface the questions worth testing. They do not see every impression or influence.

A familiar way to think about it

Marketing measurement works like a weather forecast.

Weather models disagree all the time. They can be directionally right and still miss the specifics—especially further into the future. We still check them obsessively, because more independent models and more agreement give us more confidence.

Marketing gives us one advantage: we can run tests to break ties. Experiments establish truth baselines, calibrate the models, and give a team the confidence to make the next decision.

When the reads agree, you can act with confidence. When they disagree, that's where the learning starts.

Capabilities

From the API connection to the budget recommendation.

The whole measurement stack.

Incrementality testing

Find out whether a channel created sales—or simply claimed them.

Marketing mix modeling

Understand contribution, saturation, and the strongest allocation of your next dollar.

Custom multi-touch attribution

Connect observable journeys to order records without mistaking the path for proof.

Marketing data foundations

Build the client-owned sources, definitions, and checks behind reporting and measurement.

Marketing reporting systems

Give platform, analytics, commerce, and finance numbers a clear role in the same operating view.

Which model is right?

Each one sees a different part of the weather.

That doesn't make one of them wrong. It tells you what it can—and cannot—tell you.

MMM

Broad, strategic, and inferential.

A marketing mix model sees the full portfolio over time: spend, seasonality, promotions, and demand. It is built for allocation decisions, not for reconstructing one customer journey or declaring a final truth.

Experiments

Causal, precise, and bounded.

Holdouts and controls answer the question no observational model can settle alone: what changed because we changed spend? They establish a baseline that can calibrate everything else.

MTA

Fast, granular, and incomplete.

Attribution is useful for reading observable customer behavior—especially near the bottom of the funnel. Its usefulness depends on durable first-party identity signals, and it still cannot see every impression or influence.

When the signals disagree, run a test. Marketing has one advantage over weather forecasting: we can create control groups, break ties, and use the result to calibrate the models and the clicks we're observing.

Why Brinker Analytics

I build things that stay.

Most measurement consultants interpret the data. Most data engineers build the pipelines. I do both: tests, models, warehouse, reporting, and AI-ready shared context.

When the engagement ends, your team keeps the systems, the methodology, and the ability to answer the next question without starting over.

Who I am

Eric Brinker

I'm a director-level DTC marketing analytics leader with more than a decade spent helping growth teams learn faster and spend with more confidence.

I've personally launched and run $123M in digital campaigns, led 70+ incrementality tests, and built the measurement and data systems behind high-growth brands.

Based in Nashville, I take on focused fractional engagements with one or two brands at a time.

Let's find the next useful question

Less fog. Better decisions.

Available for focused fractional engagements with one or two DTC brands at a time.

Start a conversation LinkedIn ↗