Triangulation

Analytics

Also: Measurement Triangulation · MMM Attribution Triangulation

What it isThree measurement methods cross-checked

The three sourcesMMM, attribution, incrementality tests

Why it mattersEach method alone is wrong in its own way

OutputDecisions you can defend with evidence

Quick definition

Triangulation in marketing measurement means cross-checking three independent sources of evidence: Marketing Mix Modelling (MMM), multi-touch attribution, and incrementality experiments. No single method gives the full picture. Where all three agree, you can act with confidence. Where they disagree, you have a real question worth investigating.

Try it: how attribution models redistribute credit

Paid socialMeta Ads

Paid searchGoogle Ads

EmailCRM

DirectBrand

A customer touches all four, then converts.

Credit distributed

Paid social

Paid search

60%

25%

Direct

10%

Last-click attribution buries paid social and over-credits search. The channel that closes gets the trophy.

How it varies across Australia

Triangulation is common practice at large Australian advertisers but rare below enterprise scale. Most mid-market businesses rely on attribution alone, which leaves them systematically blind to the channels attribution models undercount.

See data and tracking maturity across Australian industries →

The three methods

Marketing Mix Modelling(MMM)

Statistical regression on historical spend and sales data to estimate each channel's contribution at aggregate level.

Best for: long-run budget allocation

Multi-touch attribution(MTA)

Tracks individual customer journeys and assigns credit to touchpoints based on a chosen model.

Best for: campaign-level optimisation

Incrementality testing

Controlled holdout experiments that isolate whether a channel caused conversions that wouldn't have happened otherwise.

Best for: proving causal lift

What it actually means

Think about a navigator on a ship before GPS. They didn't rely on a single instrument. They took a compass bearing, checked the stars, and estimated speed from the propeller. Each method had its own error. But where all three pointed at the same position on the chart, they could trust it.

Marketing measurement works the same way. Marketing Mix Modelling (MMM) uses statistical regression on historical spend and revenue data to estimate each channel's contribution. Multi-touch attribution tracks individual customer journeys and assigns credit to touchpoints along the path. Incrementality experiments run controlled holdout tests to isolate whether a channel actually caused conversions that wouldn't have happened anyway.

Each method has a different blind spot. MMM is slow and can't read short-term signals well. Attribution models are fast but they confuse correlation with causation and miss touches they can't track. Incrementality tests are honest about causality but expensive to run and limited in scope.

Triangulation means running all three in parallel and looking for where the evidence converges. When MMM, attribution and an incrementality test all point at the same channel as a strong driver, you have real evidence. When one method says a channel is your top performer and the other two disagree, you have a measurement problem worth solving before you put more budget into that channel.

The businesses doing triangulation well are the ones spending the least time arguing about which dashboard is right.

One measurement method tells you a story. Three independent methods tell you what is actually happening.

How it shows up

Triangulation shows up as a structured process, not a single number. A measurement calendar with scheduled incrementality tests on major channels. An MMM model that reruns quarterly on updated spend and revenue data. Attribution reporting that sits alongside both, flagged explicitly as directional rather than causal.

The output is a channel-level view where you can see what MMM says, what attribution says, and what the most recent experiment says, side by side. Disagreements between the three become the agenda for the next planning cycle rather than the source of a slide deck argument.

Where to start if you have none of this: run one incrementality test on your highest-spend channel. The result will either confirm the attribution number or surprise you. Either outcome is more valuable than another quarter of attribution-only reporting.

The Australian context

Australian advertisers face a triangulation challenge that US advertisers don't. The smaller market means incrementality holdout tests have lower statistical power, requiring longer test windows or larger holdout groups to reach confidence. MMM models need two to three years of weekly spend data to be reliable, and many Australian businesses haven't been advertising at scale for that long.

The practical implication: Australian businesses often have to lean harder on attribution as the default and supplement with incrementality tests on their largest channels rather than running a full three-method triangulation programme. That's fine as a starting point. The risk is treating the attribution output as causal evidence when it isn't.

Where people get this wrong

Treating attribution as the ground truth and MMM as a check.Attribution is the least causally honest of the three methods. MMM and incrementality tests should be the anchor, with attribution as the operational reporting layer.

Running one incrementality test and calling it triangulation.A single experiment on one channel tells you about that channel in that window. Triangulation requires consistent, repeatable measurement across channels over time.

Using triangulation to confirm an existing view rather than challenge it.If your triangulation programme only tests channels you're confident about, it won't surface the misallocations that cost you the most. Test the channels people are most certain about first.

Common questions

Do I need all three methods to do triangulation?

Ideally yes, but start with what you have. Most businesses begin with attribution, add one incrementality test on their largest channel, then build toward MMM as they accumulate enough historical data. The value of triangulation comes from the comparison, not from any single method.

How do I handle it when MMM and attribution disagree?

Run an incrementality test on the channel they disagree about. The experiment is the tiebreaker because it is the only one of the three that tests causality directly. If the experiment sides with MMM, your attribution model has a bias you need to understand.

How much data does MMM need to be reliable?

Most practitioners recommend at least two years of weekly spend and revenue data across all channels before the model produces stable estimates. With less data, MMM outputs are directional at best and can be actively misleading. This is the biggest practical barrier for younger Australian businesses.

Is triangulation worth the effort for a mid-sized Australian business?

A full programme is expensive. A partial version is accessible: take your attribution report, run one holdout test on your top channel, and compare the results. If they agree, your attribution is working. If they disagree, you have found a misallocation worth fixing. That comparison alone justifies the effort.

Keep exploring

About New Rebellion

New Rebellion is a marketing intelligence consultancy. We build tools, score Australian businesses on how their marketing actually performs, and publish Debrief every day. This dictionary is part of how we work in the open.

How we think →