Marketing Attribution, Explained: How to Tell Which of Your Marketing Is Actually Working
Marketing attribution is how you work out which touchpoints actually drove a sale — and the default most tools use, last-click, quietly over-credits the final step and under-credits everything that created the demand in the first place.
What attribution actually means
Attribution is the practice of assigning credit for a sale across the marketing touchpoints that led to it. A customer might see a TikTok video, read a blog, get a retargeting ad, then search your name and click a Google ad before buying. Attribution decides how much of that sale each step earned. Get it wrong and you will defund the things that are quietly working.
The main models, in plain English
Last-click gives 100% of the credit to the final touch before purchase. It is the default almost everywhere because it is easy — and it is badly misleading, because the final click is usually someone already sold (a brand search, a retargeting ad) taking credit for demand created elsewhere.
First-click does the opposite, crediting the very first touch — useful for seeing what starts relationships, useless for what closes them. Linear splits credit evenly across every touch; time-decay gives more to touches nearer the sale; position-based (often 40/20/40) rewards the first and last touch most. Data-driven attribution uses your own conversion data to estimate each touch's real contribution — the most accurate, and the most demanding to run.
Why last-click quietly wastes budget
Here is the trap in practice. Brand search and retargeting always look like your best performers under last-click, because they catch people at the finish line. So teams pour money into them — and starve the upper-funnel work (content, social, PR, awareness) that actually created the demand those final clicks are harvesting. The channel that looks inefficient is often the one doing the real work.
How to measure without a data science team
You do not need a perfect model; you need to stop being fooled by last-click. Three pragmatic moves get most SMEs most of the way. First, always ask new customers how they heard about you — self-reported attribution is imperfect but catches the word-of-mouth and offline effects no tracker sees. Second, watch blended metrics: total marketing spend against total new customers, over time, across everything. If blended cost-per-acquisition falls while you invest in an unprovable channel, it is probably working.
Third, run holdout tests. Turn a channel off in one region or period and watch what happens to overall sales, not just that channel's clicks. If you pause brand search and total orders barely move, you were paying to harvest demand you would have won anyway. Geographic and on/off tests beat any attribution model because they measure incrementality — the only thing that actually matters.
The bottom line
Attribution is not about finding one true model; every model is a simplification. The goal is to stop over-trusting the last click, give credit to the work that creates demand, and use real-world tests to check what is genuinely incremental. Measure the whole journey, and you will stop quietly cutting the marketing that is keeping you in business.
Quick FAQ
What is the best marketing attribution model?
There is no single best model. Last-click is the most common and the most misleading; data-driven attribution is the most accurate but demanding. For most small businesses, combining 'how did you hear about us?' surveys, blended cost-per-acquisition and on/off holdout tests beats obsessing over a model.
Why is last-click attribution misleading?
It gives all the credit to the final touch before purchase — usually brand search or retargeting, which catch people who are already sold. That over-credits bottom-funnel channels and under-credits the awareness and content that actually created the demand.
Most businesses are not measuring their marketing — they are measuring their last click, which is a different and far more flattering thing. The channels that look inefficient are often the ones doing the heavy lifting. Ask people how they found you, watch your blended numbers, and turn things off to see what truly moves sales. — Daniel Nikolla, Founder of Merx Marketing
Stop trusting the last click — it credits the finish line and starves the work that created the demand; measure the whole journey and test what is genuinely incremental.
If you turned your best-performing channel off for a month, are you confident your total sales would actually fall?
If you would like to know more about this topic, please contact us on danieln@merxmarketing.co.uk
Sources: Marketing Minute analysis.
Related reading
Written by Daniel Nikolla, Founder of Merx Marketing Ltd and Marketing Minute
