Heat maps are cool.
Searchers’ behavior changes by hour and by day. Heat maps let you see the hot spots where you should fire up your bids — and the moments when you could save cold, hard cash by putting your ads on ice.
I’ve shared a heat map-generating script with you before, but now the tech team at Brainlabs (my employer) has improved it. In addition to looking at the overall behavior over time, you can now see performance segmented by device.
You can also get suggestions for hourly ad schedules and device bid modifiers. (Normally, you can only set six ad schedules a day and one device bid modifier for all time — but you could, say, use this handy script to change your ad schedules and mobile bid modifiers every hour.)
What does the script do? First, it averages the data, so you don’t have to worry about getting a whole number of weeks into the date range. And it smoothes the data, too, using a five-wide weighted smoothing window to get a moving average. That means you can see trends more easily, because the noise of random variation is reduced.
You can skip particular days or even set up multiple date ranges to avoid longer periods — the week around Christmas may well be anomalous, for example — but that means you may want a longer date range to make up the number of days.
But because you don’t want to be distracted by fluctuations, you need to start off with a decent amount of data. We recommend at least six weeks’ worth (as you need several instances of each day of the week) and (if you’re looking at conversion rates) long enough to get 1,500 conversions.
So, how do you turn up the heat and put your data on the map? We’ve made a template with the formatting all set up for you — the first step is to go to the File menu and make your own copy.
Then go to AdWords, set up a new Script, and copy in the code below. You’ll need to change a few settings:
["20150601, 20150930", "20160601, 20160930"]
["20151225", "20160101"]would exclude last Christmas and New Year’s Day.
["Clicks/Impressions", "Conversions/Clicks"]would give CTR and conversion rate.
["Brand", "Generic"]then only campaigns with names containing “brand” or “generic” are included. If campaignNameDoesNotContain is
["Display", "Competitor"]then any campaigns with names containing “display” or “competitor” are ignored.
Then authorize and preview the script. If it doesn’t work, check the Logs in case there was an error with any of the settings. And take care — if you run (or even preview) the script again, it will overwrite what’s already in the sheet.
Note that you’re quite likely to get fractional results, even with metrics like clicks that have to be whole numbers. This is because the heat map shows the average per hour, not the total.
If you want to tinker further, there are a couple of Advance Settings:
The post Hot off the presses: Make heat maps for device performance appeared first on Search Engine Land.