If you track your Google Ads performance month to month, you’ve probably noticed that your conversion data never seems to stabilize. You may look at last week's conversion value today and see one figure, and look again next week and see that it’s higher than before, and so on. There are many reasons you might experience discrepancies in your conversion data, but most of the time the answer is simple. What you’re probably seeing is a side effect of how Google Ads tracks conversions. There are always lags in reporting and data processing, but also, Google doesn’t track all conversions on the day they occur.
Google decided, probably 20 years ago or so, to approach this from more of an “accrual accounting” perspective and match the conversion value with the cost. The conversion is backdated from the date it actually occurred to the date of the click that its attributed to. This causes all sorts of confusion if you don’t know what’s going on, and the impact is magnified when using contemporary attribution models that can spread credit for a conversion over several different interactions in a time period that led to the ultimate sale.
So, if someone clicks on an ad today and then buys today, the conversion value is counted today. But if they clicked 3 weeks ago and didn't make a purchase, then interacted with a YouTube ad, then did another search on the brand yesterday, came back and bookmarked a product, and then returned today and made a purchase… that will be counted differently. Depending on various settings, the conversion credit may be spread across 3 different ad clicks, or it might be attributed to yesterday’s click, but it won’t show up in the reports as a sale today. The conversion value will be moved backwards in time to align with the clicks that the sale is attributed to.
The impact of this is a key reason we don’t like to look at reports right after the close of a month. Every month, Google is shifting some conversion credit to the previous period. If you pull data for the previous month sometime early the next month, you should plan to re-check that figure the following month. This will happen every time. It’s not an anomaly. It is just how Google designed the tracking system.
This often leads people to make poor decisions when reviewing short-run data because metrics like ROAS will always appear to be depressed in the past few days or couple of weeks compared to the previous period. Some amount of that always occurs, which is why we try to steer the ship with a bit longer window of data, and we don’t worry about last week’s fluctuations unless something truly dramatic occurs. Last week’s costs will be accurate, but last week’s conversion value will rise, and then rise some more, before it’s settled. You may prefer that Google didn’t do it like this, but there’s really no way around it. This is firmly entrenched in all sorts of systems, metrics, and processes now. This is just the way it is.
This is one of the key reasons to always keep this in mind: The conversion tracking system is a marketing information system, not an accounting system. It’s good for informing marketing decisions. It’s good for tracking general success metrics and steering the ship. But it was not designed to match accounting standards and will never reflect an accountant’s view of how the revenue should be tracked and realized.