For all the shiny new objects in the world of PPC, the majority of the time we PPCers spend managing and optimising accounts tends to consist of the same old activities…
Budgets and targets… Search term reports and negatives… ad performance… an occasional step up to audience, device, location, etc.
Fair enough – these are where most of the value usually lies… There’s a good reason why we settle into the 80/20 patterns that we do.
But when we’re looking to step out of the usual routine and raise the account to new heights rather than treading water, there are lots of less-visited places where you can find additional improvements with a little bit more effort.
I’ve picked out four beyond-the-usual-routine actions here.
Deeper Ad Testing
Easy to neglect (also easy to run once and then quickly neglect again!) so this is an area where we can usually do better.
Basic methods:
• Messaging emphasis (e.g. speed of service / cost / ease of use) – test 3 x pinned ads per ad group, labelled for easy comparison
• Home in on the comparative effectiveness of specific phrases. Test one phrase vs an alternative, keep all else identical, and run an Ad Variations test
• pinned vs unpinned ads (and potentially multi-pinned / part pinned… with the caveat that the pinning practice itself may be less influential than the particular pinned messages)
Advanced methods for finding winning text:
• Using either the Transparency Center or front-end searches, note repeated patterns in competitor messaging (and either seek to emulate the best OR find ways to stand out in the auction)
• Scan reviews – both your own and your competitors – and find recurring themes in what users value, or would like to see more of in the product/service you’re advertising. These become your golden benefits to emphasise. (AI is a godsend in this process)
Landing page tests (beyond 'relevance')
There’s value both in testing a complete overhaul, and – if you have sufficient volume – incremental changes to individual components.
If changes to landing pages are not available, consider whether a different existing page might be better. The text-book approach of always choosing the most ‘relevant’ page isn’t guaranteed to be the most effective.
With e-com, there are often several options in terms of how far up or down the category level you send the user.
And with local businesses in particular, the most specific service page is often not the best optimised for use as a landing page.
Campaign experiments on bid strategy
Campaign Experiments are a relatively safe way to test the effectiveness of a different bid strategy when volume is high.
Remember when running a more speculative or higher-risk test, the experiment arm doesn’t have to be set to 50%. The lower the %, the less you have to lose by testing.
PMax Image asset testing
PMax images aren’t the most obvious needle-movers. Mike Rhodes’ script and now/soon native channel reports will often tell us that PMax is barely touching the inventory that makes use of the image assets we provide… But that’s not always the case.
This is another area where AI can truly upgrade our analysis. One of the under-appreciated outputs of Mike Rhodes’ script (not yet matched by Google’s PMax improvements) is an image-by-image performance table, with all the metrics you could ask for. (NB you have to manually turn on thumbnails for its full glory.)
So if you’re stuck in ‘maintenance mode’ for a while (we’ve all been there) these next-step tests – and others like them – might be what you need to shake things up.
Dig into those deeper layers of account work – there’s often untapped value waiting there!
