Why Ad Platform Conversions Don't Show Up in Your CRM: Fixing Campaign Attribution

The ad platform says the campaign produced 40 conversions. Your CRM shows 6 deals from that campaign. Neither number is wrong; they count different things. But if budget is allocated based on one of them, every decision made without campaign attribution in place rests on a guess.

Why the two numbers never match

The ad platform counts conversions by its own definition: someone who submitted a form, clicked, or installed the app. The CRM counts customers who paid. The gap comes from three places:

Not everyone who fills a form is a fit; the filtering happens on the CRM side.
The ad platform knows the browser that saw the click, not the company that bought.
With a long sales cycle, this month's ad produces a deal three months out, and platforms do not show that lag.

Until those three gaps close, marketing and sales will keep writing different success reports for the same campaign.

The CRM fields attribution needs

Campaign attribution is not complex infrastructure; it is a handful of fields filled in with discipline:

First source: the channel the record first entered through. It must never change.
Campaign name: identical to the name in the ad platform.
First touch date and deal close date: to measure the lag.
Last touch source: the interaction immediately before the deal closed.

The immutability of first source is critical. If a rep updates the record later and first source is overwritten, all historical attribution data breaks.

Set a campaign naming rule

Naming is where attribution breaks most often. If the ad platform says "Sept-Enterprise-Test2" and the CRM says "september campaign", those two records will never match.

Pick a simple, fixed pattern: channel-period-audience-variant. As long as everyone uses the same pattern, matching becomes possible even by hand.

First touch or last touch?

Do not feel forced to pick one. Keep both and use them for different questions:

First touch shows which channel finds new customers. This is the number that defends brand awareness and top-of-funnel spend.
Last touch shows which channel closes the decision. This is the number that defends retargeting and email.

Teams that commit to one model usually cut the top of the funnel, because a last-touch model gives it no credit at all. Six months later lead volume falls and nobody understands why.

Account for the lag

In a business with a 60-day sales cycle, measuring this month's campaign against this month's deals is misleading. Shift the measurement window to match the cycle.

The practical method: report a campaign not by its launch date but by the close date of the records it produced. That way late-closing deals get credited to the right campaign.

Measure value and quality, not just volume

Leads per channel is the most misleading metric in marketing. The channel producing the cheapest leads usually has the lowest win rate.

Look at three numbers together for each channel: lead count, win rate, and average won value. If a channel brings twice the leads but wins half as often, its real cost is four times what it appears to be.

Three mistakes that break attribution

Duplicate records: the same person filling two forms creates two records and halves your conversion rate.
Manually created records: business cards from an event entered with a blank source mean that channel never appears in any report.
Overwriting the source: updating first source when a record returns erases the entire history.

The model does not have to be perfect

Many teams choose to measure nothing because they cannot build a sophisticated multi-touch model. But a rough model beats no model by a wide margin.

If the source field is filled in properly, even simple single-touch attribution largely shows where the budget should go. You can improve the model later; data you never collected cannot be recreated retroactively.

Where to start

This week's job: open the deals won last quarter and count how many have the source field populated. If it is below 70 percent, fix data capture before building an attribution model — a model cannot produce data that was never collected.

Closync matches campaign sources to won deals and shows real return by channel on a single screen.