Hi all,
I've run my fair share of Facebook campaigns over the past couple of years, and I'm about to launch a new one.
However, I'm wondering if I'm missing something. Is there a certain method to launching campaigns that allows you to scale and optimise more effectively?
E.g. how many adsets? how many ads? how much budget? etc etc.
Would be good to see if anyone has thoughts/insights!
Best,
Sam
Ha, ha. I was literally about to start a thread about this Sam. I'm in a similar situation where I've been running FB ads since around 2011 (mainly lead generation), but I don't feel like I have a good, consistent process for launching a new campaign.
I'd like to add for those who respond...
Are you split testing a landing page right out the gate with a new campaign? Or just running one?
I personally don't split test my landing page with a new campaign (I focus on finding a receptive audience), but would love to hear what other people do.
Really depends on the sort of offers you're promoting and campaigns you intend to build, FB has so many different styles of promoting and not really one method suits all.
This is my ABC Campaign structure to get the most data out of your budget:
A Better Chance (of success) Campaign Structure - For Building FB Campaigns (I couldn't think of a better name lol)
* 1 adset, multiple ads in the adset. Fb stores performance/optimization data locally at the adset level. Don't make a new adset unless you have the budget for it and it's targeting a totally different audience with zero to little overlap. An example would be testing multiple LLA.
* Start testing with images.
* Enough budget to give each ad about 8K impressions in a week. So with a 20 CPM that's $160 budgeted weekly per image tested. I like using the new FB dynamic creative feature for testing. https://www.facebook.com/business/m/...c-creative-ads
* Make your audience as big and general as possible. (Let FB and the pixel do the work of finding your ideal audience segments). If you have a list of buyers to seed a LLA use it.
* Optimize for an action that can generate at least 50 optimization events in 7 days. You may have to move up the funnel initially. Eg Add To Carts instead of Purchases.
* Try your best not to touch anything on the campaign/adset until you have your min 50 optimization events because the FB learning process resets when you make major edits
* If you want to tweak stuff, focus on optimizing and testing the LP
Mr Baffoe has some really good advice. I'm just going to pile a few things on top of what he stated.
- 1:1 image creatives are still relatively new to Facebook newsfeed ads. They're performing better than traditional 1200x628 creatives more often than not
- Don't overload your ad set with dozens of live ad variations. It will take the algorithm way longer to optimize the ad set
- Run automatic placements if you're optimizing for conversions. Let the data dictate where your ads are shown. I'm seeing a lot of advertisers turning the Stories placements off for instance, but Stories are doing very well right now. If they don't do well, Facebook won't push a lot of budget into them.
- If you're running a lookalike audience, don't narrow it. Feel free to exclude custom audiences, but I recommend not modifying the targeting settings beyond that.
- Like Mr Baffoe said, broader audiences are definitely going to be your best bet. An audience size of one million should be your absolute minimum for cold traffic (unless it's a local campaign). Go bigger than a million whenever you can.