Bidding and optimization
Start on highest volume and keep each ad set on purchases at about 50 a week. Add cost or ROAS goals only then, tied to what you can afford.
16 answers · 6 new or changed since Nov 16 · Data as of Nov 22, 2026

The short version
Our judgment, drawn from the 16 answers below.
Start on highest volume bidding, and test a cost control in a second campaign once an ad set gets about 50 conversions a week.
Anchor cost caps and ROAS goals to what you can afford, not to your average results.
Keep ad sets on purchases, and consolidate, widen the audience or raise budget before you switch to a more frequent event.
From A fallback event when purchases are scarce, Meta learning phase and limits
If your margins differ by product, send profit beside revenue and test a profit campaign against your revenue campaign.
The 16 answers
Grouped by the decision you are making. Jump to a group:
Pick the bid strategy
Volume or value, which bid strategy to start on, where to set caps and how you pay.
Which bid strategy should we start with: highest volume, cost goal or ROAS goal?
Depends on your spend. Open it to see the answer for each spend range.
Medium confidence · 9 sources
Should we optimize for the number of buyers (volume) or for revenue (value)?
Run two campaigns on the same ads, one optimized for the number of purchases and one for purchase value with a ROAS goal. Let results set the split.
Medium confidence · 8 sources
Where should we set cost caps or ROAS floors relative to current results?
Anchor caps to the cost you can afford, not your average. Set a cost per result goal about a third below your limit, and a ROAS goal close to your target.
Medium confidence · 8 sources
Should we pay Meta per impression or per link click?
Meta bills per impression by default, and ad sets optimized for purchases or leads can only be billed that way. Per-click billing comes only with a link-click goal, which buys clicks, not buyers. Optimize for purchases or leads.
New Nov 17 Medium confidence · 12 sources
Choose the event
Which event to optimize for and where your ads send people.
Which conversion event should we optimize for if we cannot reach enough purchases?
Keep ad sets on purchases at about 50 a week. Below that, consolidate, widen the audience and raise budget or cost cap. Switch to a more frequent event only if you are still short.
Medium confidence · 11 sources
If we run Meta traffic campaigns, should they optimize for landing page views or link clicks?
Optimize Traffic campaigns for landing page views. Link clicks include taps that never load your page, and Meta recommends landing page views for traffic quality. Exclude Audience Network either way, and switch to a conversion goal when you can.
New Nov 17 Medium confidence · 6 sources
Which event should app campaigns optimize for, and where should ads send people?
Optimize for purchase or subscription at about 50 a week per ad set. Short of that, merge ad sets, then use a trial or predictive in-app event. Use installs only to learn costs, and test web checkout for pricier subscriptions.
Medium confidence · 7 sources
If we sell through both our website and our own app, should Meta ads send clicks to the site, the app, or let Meta choose for each click?
Keep your website campaign as the base, and test Meta's website-and-app option as a separate campaign against it. Send app clicks with a website fallback. Move budget to whichever earns more web and app revenue.
New Nov 16 Low confidence · 6 sources
Send value and profit
Profit, lifetime value and separate targets by product or stage.
Should Meta optimize for profit or margin instead of revenue?
If your margins differ by product, send each order's profit to Meta alongside revenue, not instead of it. Then test a Purchase profit value campaign against your revenue campaign, and judge it on contribution margin after two months.
Low confidence · 8 sources
Should a subscription business optimize for first order, sign-up or lifetime value?
Optimize for first purchases, including subscription starts. Send Meta profit beside revenue or, once you qualify, a predicted lifetime value through the Conversions API. Aim for first-order break-even unless you can fund a longer payback.
Medium confidence · 12 sources
How do we build the predicted lifetime value we send to Meta?
Predict each customer's value from your own history. Confirm on past customers that it outranks first-order value and gives at least 5 distinct values, the highest at least 3 times the lowest. Then test it beside your value campaign.
Medium confidence · 8 sources
Should we set separate cost or ROAS targets by product or funnel stage?
Keep one margin-based target across acquisition. Give a product line or advertorial format its own campaign and target only when its margin or order value differs and each split still gets about 50 conversions a week.
Medium confidence · 5 sources
Fix, wait or stop
Learning, delivery problems, pausing ads and when to turn a campaign off.
How long does Meta's learning phase take, and what should we do when an ad set is stuck in Learning limited?
An ad set leaves learning after about 50 results in the week after its last significant edit, often within a week. Don't edit it meanwhile. If it shows Learning limited, combine ad sets or raise budget before changing the event.
New Nov 17 Medium confidence · 6 sources
Why is our Meta campaign not spending or not delivering, and what should we change first?
Check the Delivery column for a hard block first: a spending limit, a review or a rejection. If it spends but falls short over a full week, loosen a tight cap when costs are on target, or add new ads.
New Nov 17 Medium confidence · 6 sources
Should we pause individual ads or run automated rules on mature campaigns?
Don't switch off mature ads because past results look weak, and don't run ad-level stop-loss rules. Add new, distinct ads instead. Pull an ad when its offer ends, Meta rejects it or its rights lapse.
Low to medium confidence · 20 sources
When should we turn off a Meta campaign that isn't working, and how long do we give it first?
Write the stop rule before launch. Give a new campaign at least a week and 3 times your target cost per purchase, then turn it off if it has one purchase or none. Stop sooner only for a broken setup.
New Nov 17 Medium confidence · 8 sources
More in Meta
186 answers in 10 subjects
Guides that use these answers: How to market a lead-gen business on Meta and Google · How to start advertising on Meta