I admit, when I started doing Facebook ads 3 years ago for app installs; I made this mistake.
It cost me a couple thousand dollars and forced me to return to what I know best – Google Adwords. But since the algo update this March in Adwords; I’ve took a lot of time to invest into testing every single targeting option, ad structure, placement, trick in the book there is from a technical standpoint. I invested over $100,000 into data so it can do the talking for me.
Forget about guru products; most gurus make their money teaching you this stuff; but is what they say true or they just want it to sound good so they AWE + INSPIRE so you come back and buy their up-sell? You decide on that one.Anyways, after investing six figures into hundreds of tests ranging from $50 test, to $2k tests I’ve realized the mistake I made back 3 years ago when I was trying to get League of Legends working with Zeno’s help from STM on Facebook. The mistake was that I did not let the adsets spend enough money over a period of a minimum 3 days in order to see clear data.
I’ve invested $5 into an adset, let it run for a day, and if it didn’t bring results I culled it and got rid of it fast.
The second mistake I made was sticking to this unwritten rule you hear everywhere, that you need to test an adset with $5.
Truth be told, $5 is not much, in the USA it’s nothing because the CPM costs are high. Sure maybe in Zambia you can go a long way with $5 but data has shown me when doing testing for 1 variable, it’s wise to invest 2x the offer payout.
For example, if you are doing Ecommerce, and your margin after you subtract the cost of product is $7, it’s highly recommended to test at $15 per day, and run it for 3 days – spend $45 testing that adset.
Let’s say you have 2 angles, “LIMITED EDITION, HURRY ONLY 2 LEFT” and “SAVE $14.95! REGULAR PRICE $24.95 NOW only $9.95” and two images of a product; let’s say one where the product appears on a yellow background to signalize SALE, with a badge on it that reads LIQUIDATION SALE! and another ad image where the person is wearing your product and the background is blurred.
That means you have 2 angles, 2 ad images to test. Total 4 adsets (since we’re putting 1 image per adset)
To test this properly, you are looking at $15 per adset times 4 adsets = $60 per day * 3 (days) = $180 dollars to test properly.
TIP: If you are doing ecommerce optimize for ad to cart pixel, as you will send the FB algo data back faster so it can start sending you people who are more likely to buy vs all kinds of strangers that just like clicking on ads.
Don’t make this mistake, do it right or don’t do it at all because you will be shooting yourself in the foot!
9 thoughts on “Are you making this mistake? Facebook Ads”
Very helpfull 🙂
Great post….
As always, Great stuff and Thank You !!
Would you still use add to cart pixel for tee-spring ?
If they let you optimize add to cart, by having an event fired on the add to cart button press or after someone has added to their cart then yes.
So from your experience running ad at $5 first few days doesn’t give you any lower cpc when you start scaling?
I don’t like to start $5 adsets especially not in the USA. I want minimum 1000 impressions and the CPMs in the USA aren’t $5. They are more like $20+
Thank you for your answer and also for all helpful content which you’re sharing at this site and your fanpage.
Very, very insightful. Which marketing objective would you recommend when running mobile app install ads? especially when it comes to games