Naming conventions first, or let the AI name your campaigns?

The trouble with letting the AI pick names on its own shows up about two months later. Say a dental clinic's assistant builds six campaigns in one afternoon and calls them things like "Implants Search Broad" and "Whitening Q3 test". Nobody can quickly tell which one holds the real budget, and the account looks like three different people set it up. I'd put the naming pattern into the first prompt instead: market, channel, objective, then product, something like UK_Search_Leads_Implants. It takes five minutes, and every later report, filter and audit gets easier to read. Assistants follow a pattern well once they've seen one, so being strict up front costs almost nothing. For a throwaway test, a messy name is fine. Do you settle the naming before the first build, or rename things afterwards?

Agreed on the pattern, but I'd think about timing differently. Renaming a campaign in Google Ads doesn't reset anything, since the history sits on the campaign ID, not the name. So "Whitening Q3 test" can be fixed later without losing data. What breaks quietly is anything that matches on the name string: saved filters, Looker Studio blends, scripts or rules that look for "Search_Leads" in the title. Those stop returning rows when the name changes, and nobody notices until a report looks empty.

So settle the pattern before the first build only if you're putting reporting or automation on top of it right away. For a throwaway test, let the messy names live and rename once it becomes a real campaign. Labels can carry the things that change over time, like test versus evergreen status, so the name doesn't have to hold them.

I'm not sure "rename once it becomes a real campaign" works, because a test never really has a moment where it turns real. It just gets more budget, the messy name stays, and by the time anyone thinks about renaming it, the thing is already in a client report or a saved filter. The ID point holds, but the timing doesn't. Even a rough naming pattern for tests takes two minutes and saves you from renaming stuff other people already depend on.

Fair, a test never gets a clean switch to "real". It just quietly collects budget. So I'd skip the "rename later" plan and use the same naming pattern from the first build, even for tests, with a short prefix like TEST_ so the status is in the name. That way the name stays accurate even when nobody remembers to update it. Labels still help for filtering, but a label is one more thing someone has to check, while the name is what shows up in the client's report.

I'd push back on the TEST_ prefix, because it puts the part that changes into the part you least want to change. You said yourself that renaming breaks anything matching on the name. So a status prefix leaves you two bad options: rename a live campaign when it graduates and break the saved filters, or leave TEST_ on something that's been spending real budget for six months. Either way you're back to the "nobody remembers" problem you were trying to avoid, and the name ends up lying just like "Whitening Q3 test" did.

I'd keep the name for what never changes on a given campaign: market, channel, objective, product. Status goes in a label, or in a short tracking sheet someone updates on purpose. A label is one more thing to check, fair. But a stale label only confuses one filter, while a stale name misleads every report that uses it.

Agreed on keeping status out of the name. The campaign name is only part of what a report shows, and plenty of people filter on the campaign label or a custom column anyway. So if you do want a status label, add it to the campaign when you build it, using the same naming habit, and check it in the same audit where you look at budgets.

Elena, it depends on who reads the reports. If you're the only one in the account, a label you check during the budget audit is enough. But if a client or manager opens a default performance table, labels usually aren't in that view, so the status never reaches them. In that case a custom column, or a short sheet with campaign ID and status, is more reliable, since people tend to read whatever's on the screen they open. Either way I'd keep the name clean.

Key the sheet on campaign ID, not name. If you key on name, it breaks the same way saved filters do the moment someone renames a campaign. With the ID as the key, the status survives renames, which is the whole point of keeping the name clean.

The ID key survives renames, but copies break it. When you duplicate a campaign to test a new bid, the copy gets a brand new ID, and the sheet has no idea it exists until someone adds a row. That's the step people skip. I'd build adding the row into the copy process itself so it can't get forgotten.