CBO Optimization Cheatsheet – 5 EASY Steps for a Media Buyer to Control Campaigns
1. Check the nighttime data for the previous day – to see how over/under/close to your goal you were yesterday. That’s what you’ll build your today’s strategy around:
Check major KPIs first:
ROAS – you will usually agree with your client on the target ROAS
Amount spent – the more you spend the more you get
Cost per result–not all accounts are measured in terms of ROAS (cost/registration, cost/trial etc.)
Unique CTR and CPC – to see how good your ads are. Maybe you need new ads or different targeting
Cost/ATC, Cost/ICH – when the results are very bad but CTR and CPC are good you want to see maybe there’s a problem on the website.
2. Start turning on/off campaigns and ad sets:
– Check ALL active CBOcampaignsand ad sets first – filter out the rest. Use the following time slots:
Last 30 days – you set up goals for 30 days – you want to see how close/far from your goal you are
Last 7 days+today – turn off all campaigns and ad sets that are below your average ROAS/CPR (+-10%)
Last 3 days+today – do the same (+-20%)
– Check ALL inactive CBO campaigns and ad sets – filter out the rest. You want to find the “lazy winners” – ad sets that have become winners after Facebook has fully updated the results. Use the following time slots:
Last 7 days+today – turn on all campaigns and ad sets that are over your average ROAS/CPR (+-10%)
Last 3 days+today – do the same (+-20%)
* Adjust the time slots based on your accounts Average Order Value (AOV) and your attribution window (e.g., you should use a 14-day slot for products with AOV > $300)
3. Start running yourtests
a. Before a half of the day has passed – feel free to turn on your tests immediately. Facebook will have time to spend the money;
b. After a half of the day has passed – schedule your tests for tomorrow (either 00:01 of the accounts time or 6 AM depending on the time-zone).
– You can test the following:
New regions – except for the US, you can try United Kingdom, Australia, Canada, New Zealand, Germany, Netherlands, South Africa, Nordic countries (Sweden, Denmark, Norway, Finland), France, Spain, Israel, Switzerland, Brazil. Do not forget to include only English-speaking people in targeting.
New Lookalike audiences – best LLAs are made from the following CA:
Can be made in Manager – Pixel Purchases, FB/IG Post Engagements last 30/90d, InitiateCheckout last 30/90d, AddToCart last 30/90d, Website Traffic 180d, Video Views 95% last 30/90d
Segmented using Klaviyo or likes – 2/3+ Purchases completed, $200/300 spent, Email Subscribers list, LTV customers list (can also segment by country), 5-star reviews, VIPs (3+ purchases and $200+ spent), Churned buyers (at least 180d)
New interests – Google your competitors or large influencers, use Audience Insights, use “Suggested” interests, use “behaviors” e.g. “Engaged shoppers”.
Different placement– automatic, FB+IG feeds, FB/IG feeds+stories, FB feeds only
Manual bids– experiment with “Target Cost” and find the best spot to win auctions while paying the least amount of money
Test new creatives– if you see that “Frequency” is high or you have not changed any ads for the past few weeks;
Test different platforms– only mobile, only desktop, only iOS etc.
*Don’t be afraid to ask questions – you will usually have a supervisor or just a colleague who can give you some tips or help if anything is unclear.
* If you see that your tests immediately show poor KPIs (after ~30%*AOV spent), turn them off
4. Start optimizing:
Check your account at least once per 2-3 hours (including weekends!). Be merciless when optimizing – only your strongest audiences should pass the testing phase.
Turn off poor performers
Spent ½ of AOV with no results
Ad set’s ROAS < Average ROAS * 0.8
Scale up winners
Increase the budget by 30% – for CBOs that have a limited audience size/high ads frequency (for cold audiences – <1m in size, >1.5 in ads frequency).
Duplicate the campaign with a double budget – for larger well-established tests. Turn off the campaign with lower budget to avoid their interference on the auction.
Check KPIs – never stop checking your KPIs – look for promising targeting options and prospecting ads.
– If you have stopped many ad sets today, create new tests (and run them today if the time allows or schedule for tomorrow) – you want to keep the account’s “Amount spent” at least not less than the day before (unless the results are really bad and you need to downscale).
5. Repeat next day.
*Well, technically, you rarely repeat everything in the same manner! I mean, every campaign is unique, every test is unique, every audience behaves in a unique manner. BUT that’s exactly what I love about Facebook ads management – I learn something new every day!