The Reconciliation Gap

Every major ad platform has a version of this problem: it counts conversions using its own attribution model, inside its own data silo, with its own definition of what constitutes a "conversion." The problem is not that any single platform is lying. The problem is that when you have three platforms, a CRM, and a finance system, you get five different answers to "how many customers did we acquire last month."

The reconciliation gap is the distance between those numbers and the one that actually matters: what your system of record says you closed.

A concrete example of how this breaks

A health and wellness client came to us with solid platform performance. Meta was reporting strong ROAS. Google was showing healthy conversion volume. The management team felt good about the trajectory.

When we ran the reconciliation — pulling actual closed-won records from the CRM and matching them against ad platform conversion windows — we found the platforms had reported 7 combined conversions in the review period. The CRM showed 38 verified closed-won customers in the same window.

The platforms were undercounting by 5x.

The reason: most of the customers had converted through a longer attribution path. They saw the ad, didn't click, came back through direct or organic days later, and converted without the platform receiving credit. The platforms were optimizing toward the touchpoints they could see — and cutting spend on the channels that were actually driving closed revenue because those channels "looked" inefficient.

Once we corrected the attribution picture and gave the campaigns accurate conversion data, ROAS moved from 305% to 508%.

What to do about it

The reconciliation process is not complicated. It is just work that most operators don't do because it requires connecting systems that were never designed to talk to each other.

The minimum viable reconciliation is a monthly exercise:

1. Pull all revenue from your system of record (Shopify orders, CRM closed-won, ERP actuals) for the period 2. Pull all attributed conversions from each ad platform for the same period 3. Compare the totals and look for the gap 4. Identify the conversion paths that are falling through the platform's attribution window

From there, you can make deliberate decisions about attribution modeling rather than accepting the default.

The operators who do this consistently — even at a basic level — stop making budget decisions in the dark. The ones who don't are optimizing against platform fiction.