The branded search problem
Branded paid search is a good example. A customer sees your TV campaign. Reads about you in the press. Receives a recommendation from a friend. Searches your company name. Clicks your paid Google advert. Purchases. YAY. Paid search records a conversion. Did paid search generate the customer? Or did it simply provide the final route to a customer whose demand had already been created elsewhere? There may still be excellent reasons to bid on brand terms. But attributing the entire commercial value of that customer to paid search can materially distort investment decisions.
The affiliate problem
Affiliate can suffer from exactly the same issue. A customer decides to buy. Searches for a voucher code. Clicks an affiliate link. Converts. Last click attribution gives the affiliate the sale. Potentially overvalued. But reverse the journey. A customer discovers a product through an affiliate publisher’s gift guide. Reads the recommendation. Visits the site. Returns through Google three days later. Purchases. Google gets the sale. Affiliate gets nothing. Potentially undervalued.
The channel isn’t inherently incremental or non-incremental. The role played by the individual partner is what matters.
The CRM problem
CRM is another interesting example. A customer receives an abandoned basket email and purchases. The email platform reports the revenue. But would they have returned anyway? Alternatively, a customer receives months of useful communications, remains engaged with the brand and eventually purchases directly. CRM may receive no attribution despite potentially playing an important role in maintaining consideration.
Again: Attributed revenue and incremental revenue are not necessarily the same thing.
AI makes attribution even more complicated
AI-powered discovery introduces another largely invisible influence layer. A customer asks:
“What’s the best premium car detailing service near me?”
An AI platform recommends three businesses. The customer remembers one. Later searches for that business on Google. Clicks an organic result. Converts. Traditional analytics might report:
Organic Search generated £250
But the customer was actually introduced by an AI recommendation. As AI discovery grows, businesses will need to become increasingly comfortable with the fact that not every valuable interaction will produce a neat, trackable click.
This makes understanding the wider customer journey even more important.
So which attribution model is right?
Potentially none of them.
And all of them.
Each model provides a different lens.
The mistake is treating one model as absolute truth. At BlackBx, we are less interested in finding a single attribution model that supposedly explains everything. We’re interested in triangulating the evidence.
That means comparing:
Business revenue
Analytics data
Platform reporting
CRM data
Affiliate data
Customer acquisition
New versus returning customers
Conversion paths
Assisted interactions
Branded search
Direct traffic
Customer lifetime value
Media investment
Incrementality testing
Offline activity
Commercial performance over time
When those different signals tell broadly the same story, confidence increases. When they don’t:
That’s where we start digging.
What BlackBx seeks to decipher
Our job isn’t simply to produce an attribution dashboard. It’s to answer the questions that influence commercial decisions.
1 – CAN WE TRUST THE DATA?
Are conversions tracking correctly?
Are transactions duplicated?
Are channels missing?
Are UTMs consistent?
Is consent affecting visibility?
Are revenue values accurate?
Are platforms receiving the right signals?
2 – WHO IS CLAIMING WHAT?
How much revenue is being claimed by each platform?
Where does attribution overlap?
Do platform totals reconcile with actual business revenue?
What attribution windows are being used?
Are view-through conversions included?
3 – WHO CREATES DEMAND?
Which activity introduces new customers?
What happens to branded search when upper funnel investment changes?
Which publishers, campaigns and channels generate discovery?
4 – WHO INFLUENCES DEMAND?
Which interactions occur during consideration?
Where do customers research?
What assists conversion even when it doesn’t receive the final click?
5 – WHO CLOSES DEMAND?
Which channels appear immediately before purchase?
Are they genuinely creating additional conversions or primarily capturing existing intent?
And what is that closing role actually worth?
6 – WHAT IS INCREMENTAL?
What happens when activity is switched off? (some war stories here!)
What happens when investment increases?
Can we use geographic, audience or time-based testing?
Do sales genuinely increase or does attribution simply move between platforms?
7 – WHERE SHOULD THE NEXT £1 GO?
Ultimately, this is the question attribution needs to help us answer.
Not: “Which platform reports the best ROAS?” But:
“Where will the next pound of marketing investment generate the greatest incremental commercial return?”
That’s a much harder question and a much more valuable one.
The BlackBx view
Perfect attribution probably doesn’t exist. Modern customer journeys are too fragmented, too cross device and increasingly too privacy conscious for every interaction to be captured perfectly.
So we don’t pretend there is a magic dashboard capable of revealing the absolute truth.
Instead, we forensically interrogate the available evidence.
We validate the data. Reconcile marketing reporting against actual commercial performance. Understand attribution models and windows. Identify duplication. Separate introduction, influence and conversion. Challenge platform-reported performance. Test incrementality wherever possible and build a more credible picture of what is genuinely driving growth.
Because attribution shouldn’t be about deciding:
WHO GETS THE CREDIT?
The commercially important question is:
WHAT ACTUALLY MADE THE DIFFERENCE?
And once we understand that, we can make much better decisions about where to invest next.
BLACKBX
WE SEE WHAT OTHERS CAN’T.