Conversion fell after we turned on ads. I kept them running.
A conversion drop can accompany rising sales. Follow one investigation from the alert to the advertising change and the decision to keep traffic running.
A sharp conversion drop on an established six-pack put me into triage mode. It was one of our important products. Revenue was still up in some comparisons, but the conversion rate had moved too far to ignore.
Find where the change started
Locate the change before choosing an explanation. In this investigation, I moved from a longer view to four weeks, then fourteen days, then seven. The daily chart was hard to read, so I switched to weekly.
The latest week was better than the one before it. Useful, but not the same as being back at the old conversion level.
The counts told the real story. Units and revenue were up. Sessions were up more. That's why units per session fell. It didn't tell me where the extra traffic came from.
My first guesses were a temporary traffic spike, an email, or shoppers moving between packs. I compared the smaller pack of the same product, then the larger packs we'd recently launched. Those comparisons gave me context. The ad-history check gave me the clearest event to match against the timing.
Check the event behind the timing
October 4 was the first day we turned on ads for this six-pack.
Sessions started climbing right there: roughly 300 a week to roughly 500. Revenue went from about $12,000 a week to $13,000 around activation, then to about $16,000 in the later view.
The timing pointed to the advertising change. Sessions climbed after activation, units and revenue grew too, but sessions grew faster. That brought units per session down. The aggregate figures don’t separate how new advertising visitors converted from how existing traffic converted.
Decide whether the traffic is worth keeping
I wasn't going to pull back to restore the old conversion rate.
Revenue was at a level I hadn’t seen on this product. My decision was to keep the ads running and watch:
- Is product revenue still growing?
- Is ad efficiency slipping enough to worry about?
- Has growth settled into a new baseline that needs a different decision?
I'd still look for efficiencies inside the campaigns. That's different from reversing the traffic because one percentage fell.
| Finding | What I do with it |
|---|---|
| Sessions rise faster than units | Explain the ratio, then find the traffic source |
| Advertising starts at the onset | Check that product and timing fit |
| Revenue keeps growing | Keep running, keep watching |
| Growth stalls or efficiency drops | Reassess |
The figures above are approximate levels from the review, not a measured advertising-lift calculation. Revenue growth also needs to be judged against ad efficiency and margin. Another company with a different margin structure could reasonably make the opposite call.
Before pulling traffic back to restore an old percentage, locate the change, check the event behind it, and decide what result would justify keeping it. In this case, I wanted to see where the higher revenue settled before making the next change.