Ecommerce Attribution Beyond Last-Click: A UK Retail Guide

Most UK retailers make big budget decisions with a single, misleading number: last-click attribution. The sale gets credited to whichever channel the customer touched last — usually Google search or a discount code — while the channels that actually created the demand get nothing. The result is predictable: you cut the prospecting that fills the funnel, over-invest in the bottom of it, and wonder why growth stalls. This guide explains attribution in plain English, shows you how to set it up in GA4, and tells you what to do on Monday morning.

Why last-click lies to you

Think about how you buy things. You see a dress on Instagram, Google the brand a week later, click a shopping ad, abandon the basket, get a retargeting email, and finally buy after searching the brand name. Last-click attribution gives 100% of the credit to that final branded search. The Instagram ad that started the whole journey gets zero.

Over time, this systematically starves upper-funnel activity: social prospecting, influencer content, YouTube, display, PR. They look "unprofitable" because they rarely get the last click — even though nothing downstream would exist without them. Meanwhile branded search looks like a genius investment, when mostly it's harvesting demand other channels created.

For UK retailers this distortion is worse than average, because so much retail discovery happens on social and marketplaces while the purchase happens on Google or direct. If your reporting only shows the end of the story, you're managing the business blind.

The models, explained plainly

An attribution model is just a rule for splitting credit between the touchpoints in a customer's journey:

• Last click: all credit to the final touchpoint. Simple, standard, misleading. The default in most reporting.
• First click: all credit to the first touchpoint. Useful for seeing what introduces customers, but ignores everything that closed the sale.
• Linear: credit split equally across all touchpoints. Fairer, but treats a passing glance the same as the click that converted.
• Time decay: touchpoints closer to the purchase get more credit. Sensible for short consideration cycles.
• Position-based (U-shaped): most credit to first and last touch, less to the middle. A pragmatic compromise.
• Data-driven (DDA): Google's algorithmic model, which uses your actual conversion data to weight touchpoints. This is the default in GA4 now, and for most retailers it's the best starting point — it reflects your business, not a textbook rule.

None of these is "true". Attribution is a model, not a measurement — it's an informed estimate of influence. The goal isn't perfect truth; it's better decisions than last-click gives you.

Setting up attribution in GA4: the basics

You don't need expensive software to get started. GA4 gives you the essentials:

1. Use the Advertising snapshot. In GA4, go to Advertising → Model comparison. This lets you compare how your channels perform under different models side by side. Run last-click against data-driven and watch which channels gain credit — that's your hidden funnel.
2. Check the Conversion paths report. Advertising → Attribution → Conversion paths shows the actual sequences people take before buying: how many touchpoints, how many days, which channel combinations. Most retailers are surprised by how long and winding the paths are.
3. Set sensible lookback windows. For most retail, 30 days for acquisition events is a reasonable default; longer for high-ticket items like furniture.
4. Tag everything consistently. UTM parameters on every paid link, every email, every influencer post. Untagged traffic becomes "direct" and your model comparison is garbage. Agree a naming convention once and enforce it.
5. Import cost data. GA4 can ingest ad spend from Google Ads automatically; for Meta, TikTok and others, upload cost data or use a connector. Without costs you can compare revenue credit but not ROAS — and ROAS is the number that drives budget decisions.

One caveat: GA4's data-driven attribution only works well with volume. If your store does fewer than a few hundred conversions a month, the model has thin data to learn from — in that case, position-based is a reasonable fallback.

What to actually do on Monday morning

Theory is nice; here's the practical sequence:

1. Run the model comparison. Last-click vs data-driven, last 90 days. Identify the channels that gain the most credit under DDA — typically paid social, organic social, YouTube and display. These are your undervalued channels.
2. Check your branded search spend. If you're bidding on your own brand name in Google Ads, measure what happens when you test turning it down. Many retailers find organic captures most of it — and the budget moves to the prospecting that actually grows the customer base.
3. Re-weight, don't just re-cut. The temptation is to slash the "losers". Smarter: shift 10–20% of budget from over-credited bottom-funnel channels to the under-credited top-funnel ones, run it for a quarter, and measure total revenue — not channel ROAS in isolation.
4. Add a post-purchase survey. "How did you first hear about us?" — one question at checkout or in the order confirmation email. It's crude, it's biased, and it consistently surfaces channels your analytics misses. Cheap insurance against model blindness.
5. Review quarterly, not weekly. Attribution shifts with seasonality, promotions and creative refreshes. Weekly reactions to attribution data are noise; quarterly reviews are signal.

What attribution can't do

Honesty requires the limits too. Attribution models can't see offline influence (the friend who recommended you), PR coverage, or the podcast the customer heard. They struggle with iOS privacy restrictions and cookie loss — expect growing blind spots, not shrinking ones. And no model fixes a broken funnel: if your site converts at 0.5%, better attribution just tells you more precisely where the leaks are.

Treat attribution as one input among several: model comparison, post-purchase surveys, incrementality tests (turn a channel off in one region and watch), and plain business judgement. Retailers who combine all four make better calls than retailers who worship any one of them.

We help UK retailers get honest answers from their marketing data — GA4 setup, attribution modelling and reporting that connects spend to revenue. See our ecommerce growth service, and if paid search is part of your mix, our PPC advertising team can make sure the budget follows the evidence, not the last click. Running PPC and programmatic together? Programmatic vs PPC explains where each channel earns its place. Get a free consultation.