Been looking at hourly data on an account this week and I keep going back and fo
By Elizabeth Greene · October 3, 2026 · Curated by George's Blog
Been looking at hourly data on an account this week and I keep going back and forth on it.
Spend by hour is pretty close to real.
Clicks, mostly.
But sales attribution doesn't land at the
same speed.
So you're looking at a graph where one half of the picture is roughly current and
the other half is still filling in.
And you find a shape in it.
Spend climbs from 6pm to 9pm, sales look thin there, and now you've got a story about your evening shoppers.
Except the sales for those evening hours might show up tomorrow. Or the day after.
Day parting works.
We've shifted budget toward higher converting hours on accounts, we moved spend off weekends on one brand and total ACoS came down about three points.
But we let that data age first, and we aggregated it by week before we touched anything.
What I'm less sure about is how much of the hour-by-hour view is a reporting delay that looks like a pattern.
Still sitting with that one.