Ask ten marketers how they measure campaign performance and you'll get ten different answers — but almost all of them will mention either media mix modelling (MMM) or multi-touch attribution (MTA). In 2026, with third-party cookies finally gone from Chrome, AI-assisted modelling tools maturing fast, and CFOs demanding sharper proof of media ROI, the question of which approach to use has stopped being academic. It's a budget question.
At LUMOS, we sit in the middle of this conversation every week. Brands running programmatic DOOH alongside paid social, CTV and retail media want to know what's working — and they're tired of dashboards that don't agree with each other. So let's break down what each method actually does, where each one shines, and how Australian and New Zealand brands are increasingly combining the two.
What media mix modelling actually measures
Media mix modelling is a top-down statistical approach. It uses historical, aggregated data — typically two to three years of weekly spend, sales, pricing, distribution, weather, seasonality and macro factors — to estimate the incremental contribution of each marketing channel. MMM doesn't care about individual users. It cares about how a $100,000 shift from TV to DOOH would have moved the needle on overall sales.
MMM has had a quiet renaissance. Google's open-source Meridian library and Meta's Robyn have both seen significant updates in the past year, and the broader industry has shifted from treating MMM as a once-a-year boardroom exercise to running quarterly, semi-automated refreshes. Deloitte's 2025 CMO Survey reported MMM adoption among ANZ enterprise marketers at its highest level in over a decade, driven largely by cookie deprecation and rising privacy regulation.
What multi-touch attribution actually measures
Multi-touch attribution is a bottom-up, user-level approach. It stitches together touchpoints across a customer journey — a search ad, a retargeted social impression, an email click, a final-touch direct visit — and assigns fractional credit to each. Done well, MTA is granular, fast and tactical. You can see whether a specific creative variant on Meta drove conversions yesterday, not last quarter.
The problem? MTA was built for a cookie-era internet that no longer exists. Apple's ATT, the Chrome cookie sunset, GDPR, the Privacy Act 1988 reforms in Australia, and the rise of walled gardens have all eroded the signal MTA depends on. Most pure-play MTA vendors have quietly pivoted to 'unified measurement' or rebranded entirely. The user-level deterministic graph that powered the 2018 MTA boom is, for most brands, gone.
Strengths, weaknesses and when each one wins
The honest answer is that MMM and MTA are answering different questions, and the right tool depends on what you actually need to decide.
Choose MMM when you need to allocate budget across channels, justify a new investment to a CFO, or measure non-addressable channels like DOOH, linear TV, sponsorship and print.
Choose MTA when you need to optimise within digital channels day-to-day — creative testing, audience targeting, bid adjustments, landing page conversion paths.
MMM gives you strategic guidance with a 4-12 week lag; MTA gives you tactical signal in near real-time but only for the digital surfaces where identity still exists.
MMM handles offline media beautifully; MTA struggles with anything it can't tag.
MTA can be misleading on incrementality because it credits paths, not causation; MMM is built around incrementality from day one.
Both are improved dramatically when supplemented with geo-based incrementality tests and brand lift studies.
Why most sophisticated brands are now running both
The dichotomy of 'MMM or MTA' is increasingly a false choice. The current best practice — what Forrester and the IAB call 'triangulated measurement' or 'unified measurement' — is to run MMM as the strategic backbone, layer MTA on the digital channels where it still has signal, and use incrementality experiments (geo splits, holdout audiences, conversion lift studies) as the calibration layer that keeps both models honest.
This is particularly relevant for DOOH. A DOOH campaign cannot be measured with traditional MTA — there's no click, no cookie, no logged-in user. But MMM captures DOOH's contribution beautifully when you feed it impression-weighted exposure data, location-based audience reach, and macro-control variables. When you then run a geo-incrementality test on top — comparing exposed and unexposed regions — you get a credible, defensible read on what DOOH is actually doing to your business.
The brands winning at measurement in 2026 aren't the ones with the fanciest model — they're the ones who stopped asking 'MMM or MTA?' and started asking 'what decision am I trying to make?' That single shift in framing changes everything. — Eric Fan, CEO, LUMOS
Five questions to ask before you pick (or switch) measurement frameworks
If you're a marketing leader trying to work out what your business actually needs in the next budget cycle, start here:
What percentage of our media spend goes to non-addressable channels (DOOH, TV, audio, sponsorship)? If it's above 30%, MMM is non-negotiable.
How often do we need to reallocate budget? Quarterly favours MMM; weekly favours MTA plus experiments.
Do we have at least 24 months of clean weekly sales and spend data? If not, prioritise data foundations before any model.
Are our digital channels mostly walled-garden (Meta, TikTok, YouTube, Amazon) or open web? Walled gardens limit MTA's usefulness further.
Can we run geo-incrementality tests? If yes, both MMM and MTA become significantly more accurate. If no, that's the first capability to build.
How LUMOS approaches measurement for DOOH
At LUMOS, we don't try to sell you a measurement religion. We provide impression-grade exposure data, location-level audience reach, and movement insights that plug directly into your MMM vendor or in-house model. For brands running brand lift studies, we partner with research providers like Pureprofile to design exposed-vs-control panels. And for campaigns where incrementality is the question, we help design geo-split or daypart-split tests using our supply footprint across Australia and New Zealand.
The goal isn't to replace MMM or MTA. It's to make sure DOOH — the fastest-growing channel in Australian advertising — is properly represented in whatever measurement framework you're already running.
The bottom line for 2026
MMM is having its moment because it survived the privacy reset. MTA isn't dead, but it's a tactical tool, not a strategic one. The brands getting measurement right are the ones blending both — using MMM to set the strategy, MTA to optimise within digital, and incrementality experiments to keep everyone honest. If your current framework can't tell you what DOOH, CTV and retail media are actually contributing to revenue, it's time for an upgrade.
If you'd like to talk through how DOOH fits into your measurement stack — whether you're running MMM, MTA, or piecing something together — the LUMOS team is happy to share what we're seeing across the AU/NZ market. Reach out via spotlumos.com.
[SCHEMA_ORG]{"@context":"https://schema.org","@type":"Article","headline":"Media Mix Modelling vs Multi-Touch Attribution: Which Does Your Brand Actually Need?","description":"Media Mix Modelling vs Multi-Touch Attribution: Which Does Your Brand Actually Need?","image":"https://images.unsplash.com/photo-1559136555-9303baea8ebd?w=1200&q=80&fm=jpg","datePublished":"2026-06-05","dateModified":"2026-06-05","author":{"@type":"Organization","name":"Lumos","url":"https://spotlumos.com"},"publisher":{"@type":"Organization","name":"Lumos","url":"https://spotlumos.com","logo":{"@type":"ImageObject","url":"https://spotlumos.com/logo.png"}},"mainEntityOfPage":{"@type":"WebPage","@id":"https://spotlumos.com/au/resources/media-mix-modelling-vs-multi-touch-attribution-2026"},"keywords":"programmatic DOOH, digital out-of-home, audience intelligence, media planning","about":{"@type":"Thing","name":"Programmatic DOOH Advertising"},"isPartOf":{"@type":"WebSite","name":"Lumos","url":"https://spotlumos.com"}}
