-
At what point do you have a statistically signifcant amount of data to narrow down age targetting?
Posted by on November 9, 2020 at 9:23 pmHey folks! Here’s the deal, I have a campaign that has gotten me 10 purchases all from women. I have a reach of >1500 in each age range. All of the age ranges have 1 purchase, except for the 45-54 age range which has 5 purchases. Obviously 45-54 has the lowest cost per purchase. Can I confidently narrow down to this age range? I checked and it shows I still would reach 320,000 within this age constraint. Thanks in advance.
-
3 Replies
-
That’s a loaded question. Statistical significance is decided based on effect size you’re looking for, sample size, general population, & the model you want to use (chi-sqaured, t-test, etc.) which in turn depends on the kind of data you’re comparing.
Based on how your question is worded, I recommend some light reading to start and only work your way up to ‘statistical significance’ once you’re sure it’s applicable…
[https://www.searchenginejournal.com/interpret-ppc-data-correctly/371937/#close](https://www.searchenginejournal.com/interpret-ppc-data-correctly/371937/#close) -
In my experience and boatload more data, if your talking white and blue social media, I would bet the farm that if you went 45-54W, it would see a ROAS of .025. HA.. The more data driven and narrower the audience, the worse the results I’ve seen for whatever reason–especially in last 60 days. Literally have thousands of conversions from various demographics and when a campaign is segmented off to hunt the best ROAS age segment, it goes to complete hell.
I sincerely hope its the opposite for you but If anything, I’d keep what you have going and spin a duplicate, narrowing down to that single segment to see how it goes. Your overlap will be minimal and if you don’t have a massive budget, it will be noise in terms of bidding against yourself. 320K audience isn’t tiny, but it’s not a lot unless your budget is relatively small (under $100 a day) IMO. If you budget is relatively big, I’d try to tinker with min ROAS cap to avoid getting shellacked–though you may see that spend suffers.
Your best data is going to be experimenting and figuring out when to double/triple down when things are working well. I’ve found that with all the statistical data in the world, virtually all online advertising platforms are wanting to dummy down the entry (any mom and pop an advertise for themselves) and are taking control with “smart” bidding. Translated, crap loads of data and very strategic bid tiers ain’t what it used to be IME.
-
Age bid adjustments aren’t really the most accurate, unless you have a huge volume of traffic and conversions. 10 purchases in nowhere near that. You can up the bids on female demographics but don’t lower it on male demographics.
Log in to reply.