Why do Meta and Google both report the same conversion?

Because each platform counts any conversion it can connect to an interaction it served, within its own attribution window. Both genuinely touched the customer. Neither can see the other. Adding their numbers together always overstates reality.

Which attribution model should we use?

For most Indian service businesses, something that spreads credit across the journey rather than giving it all to the final click. But the model matters less than picking one and staying consistent — switching models changes every historical number and makes comparison impossible.

Is "how did you hear about us?" data reliable?

Not precisely. People misremember and pick the most recent thing. But it is often the only signal that captures activations, word of mouth and signage at all — channels every digital model scores as zero. Treat it as a rough weighting, not a measurement.

How do we measure offline marketing's contribution to online leads?

Rarely cleanly. The workable approaches are geographic — comparing enquiry volume in areas where you ran activity against areas where you did not — and temporal, watching whether branded search and direct traffic rise after offline activity. Both are crude. Both beat assuming the contribution is zero because it is hard to see.

Performance Marketing

Who gets the credit for the sale?

Add up what every platform claims it delivered and you will have sold roughly twice as much as you actually did. All of them are telling the truth as they see it.

31 August 20266 min read

The double-counting nobody mentions

Meta reports the conversion. Google reports the same conversion. Your analytics reports it once, credited to whichever came last.

Nobody is lying. Each platform can only see the part it touched, and each is built to report on itself. A person who saw a Meta ad on Tuesday, searched your name on Thursday and bought on Friday genuinely interacted with both — and both will count it.

Sum the platform dashboards and you get a number bigger than your actual sales. This is normal, and it is the first thing to understand before making any budget decision from those reports.

Why the last click is the worst judge

Default reporting usually gives full credit to the final interaction. It is the easiest thing to measure and the least responsible for the outcome.

The last click is almost always the cheapest and most efficient-looking one — often a brand-name search from somebody who had already decided. Something made them decide. That something rarely gets the credit, because it happened earlier and less measurably.

Judge budget on last-click alone and you will steadily move money toward the bottom of the funnel — buying more of the clicks from people who were going to buy anyway, and less of whatever created the intent. It looks efficient for a quarter, then the pipeline thins, and the report cannot explain why.

The part no attribution model sees at all

Every attribution model, however sophisticated, only measures what happened on a screen.

It cannot see the activation somebody attended. It cannot see the hoarding on a daily commute, the dealer who recommended you, or the fact that a family has been passing your board for four years and therefore feels safe buying from you.

For a business like this, where a great deal of demand is created offline and only *captured* online, a purely digital attribution report is not measuring your marketing. It is measuring the last few metres of it.

What to actually do about it

Nobody solves attribution. The practical answers are less elegant and more useful:

Pick one source of truth and stay with it. Usually your own analytics or CRM, not the ad platforms. It will undercount some channels — but it is consistent, and consistency is what lets you see whether things are improving.

Read platform numbers as directional, not additive. Meta's own reporting is useful for comparing Meta campaigns against each other. It is not useful for deciding what share of your total budget Meta deserves.

Ask the buyer. A single question on the enquiry form — "how did you hear about us?" — is imprecise, self-reported, and often the only thing that captures the offline half at all. Imperfect data about the invisible channel beats perfect data about the visible one.

Watch the totals, not the splits. When you increase spend somewhere, do total enquiries rise? That question is harder to argue with than any attribution model, and it is the one that actually answers the budget question.

The honest position

Attribution is directionally useful and precisely wrong. Treat a dashboard's confident percentage as an opinion with a methodology, not a fact.

The teams that get this right are the ones comfortable saying: this channel probably contributed, we cannot separate it cleanly, and here is what we watched instead. That reads as less rigorous than a pie chart. It is considerably more rigorous. It is the position to take in every digital marketing review, and the one clients thank you for a year later.

Questions people ask

Why do Meta and Google both report the same conversion?: Because each platform counts any conversion it can connect to an interaction it served, within its own attribution window. Both genuinely touched the customer. Neither can see the other. Adding their numbers together always overstates reality.
Which attribution model should we use?: For most Indian service businesses, something that spreads credit across the journey rather than giving it all to the final click. But the model matters less than picking one and staying consistent — switching models changes every historical number and makes comparison impossible.
Is "how did you hear about us?" data reliable?: Not precisely. People misremember and pick the most recent thing. But it is often the only signal that captures activations, word of mouth and signage at all — channels every digital model scores as zero. Treat it as a rough weighting, not a measurement.
How do we measure offline marketing's contribution to online leads?: Rarely cleanly. The workable approaches are geographic — comparing enquiry volume in areas where you ran activity against areas where you did not — and temporal, watching whether branded search and direct traffic rise after offline activity. Both are crude. Both beat assuming the contribution is zero because it is hard to see.