We love talking about metrics that are easy to get. Digital media made that the norm – CPMs, CPCs, CTRs, CVRs, CPAs, all instant. Trouble is, they can’t all go up at once. That’s not how digital marketing works. So we settle for a report full of green, amber and red, and call it effectiveness.
The problem isn’t that these numbers are wrong. It’s that they were never built to carry the weight we’ve put on them.
Vanity metrics wearing a suit
ROAS looks rigorous because it has a currency sign in front of it, but a single-channel ROAS figure with no view of incrementality, brand contribution, or what happens three months later is not rigour. It’s a number that’s easy to put in a QBR.
The problem with defaulting to single platform reporting as a “single source of truth” – usually last-click attribution linked to web analytics – is predictable: paid search carries credit for demand it never generated, sitting in the last touchpoint and hoovering up credit other activity earned upstream. So many brands are still running on that single lens (with management asking “how can we spend more in PPC Brand?”) and calling it measurement.
None of this is anyone being lazy. Single metrics are fast to report and simple to attach a target to. Multi-lens measurement is slower and harder to defend in a five-minute update – so the organisation reaches for the version of “effectiveness” that fits the meeting, not the one that reflects reality. It also can’t see what’s driving demand into the channels it credits – paid search, organic, AI search capture demand, they don’t create it, and what creates it is rarely the same channel twice, which is why you need a proper test and learn programme, not a template.
Multiple lenses, not one
Effectiveness isn’t one measurement. It’s a triangulation of several imperfect ones, none of which is sufficient alone:
- Short-term response: did the activity do what it was bought to do, in the window it was bought to do it in.
- Incrementality: what would have happened anyway? The lens most businesses skip, because it’s the one that most often deflates the headline number.
- Brand contribution: the effect that shows up in consideration, in search, in the performance of every other channel weeks later. Slower, harder to attribute, still real.
- Diminishing returns and saturation: is the channel still earning its marginal spend, or living off historic efficiency that’s quietly eroding?
- Quality of the outcome, not just the volume of it: a lead is not just a lead. Some of the cheapest conversions are the least valuable ones sitting on the balance sheet a year later.
No single approach gives you everything – hence triangulating digital attribution, marketing mix modelling and controlled experiments. At HSBC we called the adjustment factor an “inflator”; at Lloyds it’s “modifier” – same job, correcting a platform number back to what incrementality evidence supports. MMM is no longer just for seven-figure budgets: Google and Meta offer free tooling, which rightly makes people wary of the media owner marking its own homework – the fix is clean data and more than one model. Tools like Sellforte are pushing it towards real time.
Where measurement should sit
Rather than pouring more into media, or grander creative and more influencers, isn’t it worth “investing” in effectiveness first, so existing plans work harder? Attribution alone is complicated; a custom model that’s right for your business is another order of difficulty, and most businesses skip asking if they have the resources for it.
Ownership matters too. At Amazon, Data Science sat under Finance, taking the politics out of evaluating performance. Elsewhere, Marketing marks its own homework, and spends the year teaching everyone else why its numbers are the good story they look like.
The same logic decides which metric belongs where. Incrementality suits finance, or quarterly business contribution – slow, expensive, answering whether spend is genuinely adding to the business. Not a Tuesday question. Day to day with an agency, gross conversions are the right tool – fast, visible to both sides. Insist on incrementality-grade rigour daily and the team reverts to gut feel. The skill is matching the metric to the altitude, not picking a “correct” one.
The uncomfortable part
Multi-lens measurement will, sooner or later, contradict a number someone has already put their name against – telling you the channel that looks brilliant on ROAS is cannibalising brand search. Not popular where bonuses ride on the simpler story, but effectiveness done properly will occasionally make you the person who ruins a good news story with a better question. That’s the job.
If I were fixing this tomorrow, I’d ask every channel owner one thing: what would you have to see to admit this isn’t working? If nobody can answer that, you have a scoreboard that only ever counts one way.
Jay Safdar is a growth marketing leader with experience spanning global financial giants, British icons and tech innovators – including AllSaints, Amex, Amazon, Barclays, EE, Goldman Sachs, HSBC, Hyperoptic, John Lewis, Lloyds and WEX Inc. Specialising in demand generation, performance marketing and marketing effectiveness, Jay has a proven track record of driving commercial impact through data-led strategy and insight-driven creative.
