Which attribution approach should a consumer brand our size use?
Our answer
Use store or CRM totals and total sales divided by total ad spend (MER) for the business, platform reports within platforms and incrementality tests for budget between channels. Add paid attribution at about $50K a month, a mix model later.
- Confidence
- Medium
- Strongest proof
- 1 official rule
- 3 datasets, 9 practitioners, 6 opinions or claims
- Sources
- 18 sources
- 19 statements
New here? Take the 1-minute tour
Why this is the answer
3 reasons and our judgmentEach decision needs a different tool. Grüns uses tests and a mix model for big budget calls and attribution for day-to-day moves. Silverback Strategies argues that most brands are overinvested in attribution and underinvested in incrementality testing.
Most large stores run without a paid attribution tool. Eightx found about 87% of 74,777 Shopify Plus stores run no dedicated attribution platform. Jetfuel Agency says paid tools pay off from $50K a month across channels.
Mix models are built to lean on experiments. Google and Meta build their open-source mix models to be calibrated with experiments. Recast, which sells one, says most businesses do not need a mix model.
Our judgment
Give each method one job: store totals and MER for the business, platform reporting inside each platform and tests for budget between channels. We take this from Grüns's practice, Silverback Strategies' guide and QRY's measurement plan. Add paid tools in order of spend, attribution at about $50K a month across several channels and a mix model only with 3 or more channels and the weekly history Whether you need a media mix model sets. That order rests on Jetfuel Agency's tool threshold, Flighted's and Eightx's channel counts and the data minimums of Eightx (18 to 24 months) and Balistro (12 to 18 months). HexClad's Northbeam endorsement doesn't change the answer, because HexClad runs the tool beside Haus geo holdouts and a mix model, which is still a layered setup.
Our judgment, built from Grüns, Silverback Strategies, QRY, Jetfuel, Flighted, Eightx, HexClad and Balistro: our reading of these sources, not a finding from any one of them.
What the sources say
19 statements from 18 sourcesIn a Funnel-sponsored survey of 233 marketers, only 9% combined incrementality, mix modeling and attribution, and 61% chose a metric because it told a better story. Fieldwork details were not disclosed.
“36% of in-house marketers kept funding a channel they believed was not working because its numbers looked too good to cut.”
A Cornell Tech paper shows that because ad algorithms target deterministically, observational ad measurement can be substantially biased, and it calls a fully randomized experiment the ideal.
“the violation of overlap can substantially bias the estimates of the average treatment effect from observational data”
Meta's Robyn mix model docs say ad returns diminish, so budgets should follow the next dollar's response, and models should be calibrated with lift or geo tests. A relayed whitepaper puts uncalibrated models 25% off.
“uncalibrated models show 25% average difference to the ground truth”
Gruns' growth head matches the tool to the decision: incrementality tests, a mix model and marginal return for big budget calls, and attribution and platform data for daily moves.
“I'm also not letting attribution decide whether Meta deserves another $500K.”
Zigpoll matched 3.3 million post-purchase survey answers from 3,237 stores to tracked sources: 51.8% of orders had no tracked source. YouTube, word of mouth and TikTok sales went untracked most; Facebook least.
“51.8% of orders arrived with no usable source.”
Google says there is no single formula for turning an experiment into a mix model prior: combine evidence and allow extra uncertainty, since lift tests and the model measure different things.
“There is no single formula to translate an experiment result into a prior.”
Cody Plofker's first independent check on TV is the post-purchase survey, scaled up by the response rate (if half of buyers answer, double it), then compared with other attribution models.
“never allow anyone to create their own homework”
Jetfuel, an agency, says GA4 attribution is usually enough under $20K a month, and third-party tools are worth it at $50,000+ a month across several channels. It gives no data.
“Worth the cost when you are spending $50,000 or more per month across multiple channels”
HexClad uses Northbeam to compare channels on one basis regardless of each platform's attribution, runs geo holdout tests with Haus, and keeps a mix model.
“the backbone of our media buying decisions”
Michael Kaminsky, whose firm Recast sells mix models, says most businesses do not need one and it can mislead them; he cites studies where last-touch was about 85 percent as good as experiments in simpler mixes.
“most businesses do not need an MMM. In fact, an MMM will be net negative for them because it will actively mislead them.”
Flighted says a mix model fits brands spending $50K+ a month across 3 or more channels, while Meta-only brands with same-day purchases do better with incrementality tests.
“spending $50K+/month across channels”
Balistro, an agency, says to add up the conversions every platform claims and compare with real orders; for most DTC brands it audits, the sum is 1.4x to 2x real orders.
“For most D2C brands we audit, the platform sum is 1.4x to 2x real orders.”
Silverback Strategies, which sells media and testing, says most brands overinvest in attribution and underinvest in incrementality: keep attribution for daily optimizing and test your biggest channel.
“Most brands are overinvested in attribution and underinvested in incrementality testing.”
Skullcandy's ecommerce head says a sophisticated multi-touch model is not the answer; he uses incrementality studies and an attribution layer to make the business case to finance.
“I don't think the solve is having some sophisticated like multi-touch attribution model”
Eightx found about 87% of 74,777 Shopify Plus stores run no dedicated attribution platform (Triple Whale is on 11.42%), and says to fix tracking before buying one.
“Fix the wiring before you sign the contract”
An Operators podcast host says a brand doing about $5M a year that runs only conversion-optimized Meta campaigns needs neither incrementality testing nor multi-touch attribution tools.
“You don't need incrementality and you do not need MTA tools.”
Eightx puts the mix model adoption floor near $20M+ revenue and $3M to $5M+ a year in media across 3 or more channels; under $1M, use platform reports plus quarterly incrementality tests.
“The adoption floor for DTC is roughly $20M+ in revenue and $3M-$5M+ in annual media spend”
CTC, long critical of multi-touch attribution, now offers Northbeam to smaller brands, since at $10K to $15K a month a store likely lacks the conversions per region for geo tests.
QRY, an agency, maps tools to annual media spend: under $1M Triple Whale or Northbeam; $1M-$3M Northbeam or Rockerbox plus Haus tests; $3M+ add Measured or Recast. It says to write a measurement plan first.
“$1M to $3M: Northbeam or Rockerbox for channel attribution; Haus for incrementality testing on your top channels.”
In the full answer
For readers who sign in- What to do this week3 steps
- The numbers behind it3 figures from cited sources
- Where each source standsA chart placing each source on the question
- The strongest disagreementThe best case against the answer, from a cited source
- How sure we areWhat its confidence rests on, and what would change the answer