Quick answer: what should a Meta ads optimization checklist include?
A Meta ads optimization checklist splits work by cadence. Daily, you watch spend, tracking, rejected ads and delivery status without editing. Weekly, you make batched decisions using spend-based kill rules and fatigue signals. Monthly, you audit attribution, tax, break-even ROAS on delivered orders, structure and billing. Most days should end with no edits.
How to use this checklist: watch daily, decide weekly, restructure monthly
The routine that works at ₹10 lakh to ₹1 crore a month separates watching, deciding and restructuring. You watch every day, make changes once a week in one batch, and review measurement and structure once a month.
The reason is the learning phase. Meta's Help Center defines a significant edit as pausing an ad set or changing its optimisation event, audience or creative. Bid strategy and budget changes may also count, depending on how big they are.
Each significant edit restarts the count of about 50 optimisation events an ad set needs in the following week to exit learning. A team that edits daily keeps sending that count back to zero.
This page covers the routine only. For the strategy behind it, see our complete 2026 guide to Meta ads optimization.
Set up three things before day 1
- A saved column preset in Ads Manager: amount spent, results, cost per result, purchase ROAS (return on ad spend), CTR (click-through rate), CPM (cost per 1,000 impressions), frequency, delivery and Last Significant Edit.
- A written change log. Use a shared sheet where every edit gets a date, an ad set and a reason. There is a template at the end of this page.
- Your target CPA (cost per acquisition) and break-even ROAS, calculated on delivered, tax-exclusive revenue. The next section shows how.
On ownership, the founder or marketing head reads the daily summary. The media buyer, in-house or at the agency, runs the weekly review. Both sit together for the monthly audit, ideally with someone from finance.
Set your targets first: break-even ROAS on delivered, tax-exclusive revenue
Every threshold in this checklist compares results with a target, so the target must reflect real profit. Calculate it on tax-exclusive revenue from delivered orders, not on the purchase value Meta shows.
Meta's purchase value comes from your store. It often includes GST in India, VAT in the UK, EU and Gulf, or sales tax in the US. In India, the Gulf and parts of South-East Asia, it also includes cash on delivery (COD) orders that are refused at the door. These return-to-origin (RTO) orders show as purchases in Meta but never pay you.
Turning unit costs into a break-even ROAS
You need four inputs: tax-exclusive order value, product cost, shipping and fees, and the share of orders actually delivered. Our break-even ROAS guide with formulas and a calculator covers the full method.
Worked example: a brand spending ₹30 lakh a month in India
Step 1: Find profit per placed order before ads
Example: the average order is ₹2,360 including 18% GST, which is ₹2,000 excluding tax. Product cost is ₹600, shipping ₹150, and payment and COD fees ₹50. That leaves ₹1,200 per delivered order.
80% of orders are delivered. Each RTO costs about ₹250 in two-way shipping.
Profit per placed order
(0.8 × ₹1,200) - (0.2 × ₹250) = ₹960 - ₹50 = ₹910Step 2: Convert it into the ROAS Meta will show
₹910 is your break-even CPA. Meta reports ₹2,360 of revenue per purchase, so break-even on Meta's own numbers is higher than most teams assume.
Break-even ROAS on Meta-reported revenue
₹2,360 ÷ ₹910 = 2.59xStep 3: Adjust for tax on ad spend
According to industry and tax sources, Meta's India entity adds 18% GST to the ad bill, and with your GSTIN (GST registration number) on the account you can usually claim it back as input tax credit (ITC). Confirm this with your accountant. Without it, every ₹1 lakh of spend really costs ₹1,18,000, and break-even rises to about 3.06x (2.59 × 1.18).
Our rule of thumb is to set the target CPA at roughly 75-80% of break-even. Here that is about ₹700.
In this example, a campaign showing 2.5x ROAS in Ads Manager is losing money on every order. More on this in why a 3x ROAS campaign can still lose money.
Daily checklist: 10-15 minutes, mostly no edits
The daily check should end with no edits on most days. You are looking for broken things, not for weak ads.
- Spend pacing: a ₹30 lakh month is about ₹1 lakh a day. Flag any day that lands far from plan, such as ₹1,40,000 or ₹60,000, and find out why.
- Store orders vs Meta purchases: learn your normal ratio over a few weeks, then flag days well outside it. If Meta purchases collapse while store orders hold, tracking probably broke. If store orders fall while clicks hold, check checkout and payments, including UPI failures in India.
- Pixel and CAPI heartbeat: in Events Manager, confirm that purchase events are still arriving from both the browser pixel and the Conversions API (CAPI). See our Pixel and Conversions API setup guide if either looks off.
- Rejected ads, Account Quality and payment failures: these stop delivery silently, so check them every day.
- Delivery column: note any ad sets in Learning or Learning Limited and any ads marked Creative Limited or Creative Fatigue. Write a note only, don't act yet.
- Early fatigue signals: compare frequency, CTR and hook rate (3-second video views ÷ impressions) on your top-spending ads with each ad's own first week.
Read only data since the last significant edit
Meta's help page says to set your reporting range from the last significant edit. If the edit was 7 days ago, look at the last 7 days. Ads Manager has a "since last significant edit" date preset for this, according to several practitioner guides (we could not confirm the preset name in Meta's Help Center).
If an ad set has had fewer than about 50 optimisation events since that edit, treat its cost per result as noise.
There are only three: fix broken tracking, fix policy or payment problems, and stop runaway spend. Everything else goes into the change log for the weekly review.
A short daily summary of pacing and the store-vs-Meta gap lets a founder open Ads Manager only when something is actually off.
Weekly checklist: 45-90 minutes, one batch of decisions
The weekly review is where every change happens. Make the changes in one batch, on one day, and log each one with its date.
- Run breakdowns over 7-10 days by placement, region or state, age and gender, and device. Cut only clear losers that have had real spend. Our 7-day breakdown method shows how to read them.
- Apply the kill rule at ad level. A common practitioner rule of thumb (directional, not a Meta rule): pause an ad if its CPA is more than 20% above target after it has spent 2-3x target CPA. Also pause it if it has spent 1x target CPA with zero purchases. With a ₹700 target, that means pausing at a CPA above ₹840 after ₹1,400-₹2,100 of spend, or at ₹700 spent with no sale.
- Confirm fatigue before replacing creative. Look for frequency rising past about 2.0-2.5 for ecommerce (around 3.0 for broad prospecting; both are practitioner rules of thumb) together with falling CTR or hook rate. High frequency with stable CTR and conversion rate is not fatigue. Our creative fatigue guide covers the fixes.
- Launch 2-3 new concepts on the same day, so each ad set takes one significant edit instead of several spread across the week.
- Move budget in one change per ad set, and log the date you will re-read the results.
- Deal with Learning Limited ad sets. Consolidate them, or fund them toward 50 events a week at target CPA. At ₹700, that is about ₹35,000 a week, or ₹5,000 a day.
Meta's Help Centre says Creative limited appears when cost per result is higher than your past ads but less than twice as much. Creative fatigue appears when it is twice as much or more. By the time either label appears, you have already paid for the decline, so use them to confirm what your weekly trend check found.
Avoid switching the optimisation event to add-to-cart just to escape Learning Limited. For ecommerce, it usually brings in lower-quality buyers.
Monthly checklist: half a day on measurement, money and structure
The monthly audit checks that the numbers you act on are still true. Most of it is measurement, not media buying.
Measurement audit
- Removed attribution windows: In a developer blog post dated 16 October 2025, Meta said that from 12 January 2026 the Ads Insights API no longer returns the 7-day view and 28-day view attribution windows. According to PPC Land's report, these requests return empty data, not an error. Check every dashboard and connector, and re-base any comparison that crosses January 2026.
- Engage-through: On 3 March 2026 Meta said click-through attribution for website and in-store conversions will count link clicks only. Conversions that follow a share, a save or another non-link click move to engaged-view attribution, which Meta is renaming engage-through. Add a note to monthly reports, because click-attributed numbers shifted.
- Data retention: The same Meta post limits frequency breakdowns to 6 months of history, and hourly breakdowns and breakdowns of unique-count fields to 13 months. Totals stay available for up to 37 months. Export any breakdown you want to keep for longer.
- Signal quality: check the Event Match Quality (EMQ) trend and CAPI deduplication. Where your setup allows, send purchase value excluding tax and shipping.
Money, structure and planning
- Recalculate break-even ROAS using last month's delivered rate and return rate.
- Merge overlapping audiences, archive dead campaigns and cut ad sets your budget cannot feed.
- Confirm your tax ID is on the account, send invoices to finance, review admin and partner access, and keep a backup payment method.
- Plan next month's creative pipeline and seasonal budgets. Plan a holdout or incrementality test each quarter, so you are not relying only on Meta-reported ROAS.
Festive periods such as Diwali, Ramadan and BFCM (Black Friday and Cyber Monday) push CPMs up. Expect it in your plan rather than reacting with daily edits.
Market-specific adjustments for tax, COD and RTO
The principle is the same everywhere: judge ROAS on tax-exclusive, delivered revenue. What changes by market is how tax appears on the Meta bill and how much COD you carry.
| Market | Tax on the Meta bill | What to set up | ROAS note |
|---|---|---|---|
| India | 18% GST (industry sources say India accounts have been billed by Facebook India Online Services since October 2018; not confirmed from Meta) | Add GSTIN to claim ITC. Tax advisers say that if you are invoiced by Meta Platforms Ireland, you self-assess 18% IGST under reverse charge; confirm with your accountant | COD and RTO are common, so count delivered orders only |
| Australia | 10% GST added by default, according to industry sources citing Meta's Australian GST help page | Add a valid ABN and confirm GST registration | Strip GST from revenue |
| UK and EU | Often invoiced by Meta Platforms Ireland, according to industry sources | Set business and billing country correctly. Reverse charge usually applies to VAT-registered businesses in other EU states | Strip VAT from revenue |
| Gulf, US, South-East Asia | Varies (VAT in the Gulf, sales tax in the US) | Confirm local rules with your finance team | COD is common in the Gulf and parts of South-East Asia |
A common mistake is comparing ad spend that includes tax with revenue that excludes it, or the reverse. Pick one basis and keep it in every report.
What most guides get wrong about Meta ads checklists
Most ranking checklists repeat rules that are outdated or that Meta never published. Here are the ones to drop.
- "Leave learning alone for 7-10 days." Meta defines learning by volume, about 50 optimisation events since the last significant edit. An under-funded ad set stays Learning Limited however long you wait.
- "Budget changes over 20% reset learning." Meta only says budget and bid changes may count, depending on size. 20% is a sensible practitioner rule of thumb, not Meta's rule.
- "Cap frequency at 5." Meta publishes no frequency threshold. Practitioners often see ecommerce fatigue start around 2.0-2.5. Meta's own fatigue status looks at how often people have seen the ad and at its cost per result, not at a fixed frequency number.
- "Adjust bid strategy daily." A bid strategy change can restart learning, so it belongs in the weekly batch.
- "Kill after 3-5 conversions." That is far too little data. Judge ads by spend relative to target CPA instead.
- Dashboards still showing 7-day view numbers. Meta's API stopped returning the 7-day view and 28-day view windows on 12 January 2026, so those columns are now blank.
Thresholds cheat sheet and change log template
Print this and keep it next to Ads Manager. All thresholds are practitioner rules of thumb, not Meta rules, so adjust them to your margins.
| Check | Cadence | Trigger | Action |
|---|---|---|---|
| Spend pacing | Daily | Day far off plan | Find the cause and fix runaway spend only |
| Store vs Meta purchases | Daily | Ratio well outside normal | Check pixel, CAPI and checkout |
| Ad CPA | Weekly | Over 20% above target after 2-3x target CPA spent | Pause in the weekly batch |
| Zero-sale ad | Weekly | 1x target CPA spent, no purchase | Pause |
| Fatigue | Weekly | Frequency above 2.0-2.5 plus falling CTR or hook rate | Rotate in new concepts |
| Learning Limited | Weekly | Under ~50 events a week | Consolidate or fund it |
| Attribution and tax | Monthly | Blank fields, missing tax ID | Fix connectors and billing |
Change log columns
- Date and ad set.
- Edit type: significant or not.
- What changed and why.
- The date to re-read results, once about 50 events have landed.
The change log is what makes weekly decisions possible. Without it, nobody knows which results come after an edit and which come before.
Read next in this series
- Meta Ads Optimization: The Complete 2026 Guide
- Meta Ads Optimization With AI: The 7-Day Breakdown Method
- How to Reduce Cost per Purchase on Meta Ads
- Meta Ads Learning Phase: How to Exit It Faster
- Meta Ads Placement, Age and Region Breakdowns: What to Cut
- Meta Ads Optimization for ₹10-20 Lakh Monthly Budgets
- Advantage+ Shopping Campaign Optimization Guide
- Why Your Meta Ads Stopped Converting (and How to Fix It)
- Meta Ads Optimization for Lead Generation
Key takeaways
- Watch daily, decide weekly, restructure monthly. Each significant edit restarts the roughly 50-event count an ad set needs to exit learning.
- Read results only since the Last Significant Edit. Ignore ad sets with fewer than about 50 optimisation events since that date.
- Judge every ad against a break-even ROAS built on delivered, tax-exclusive revenue, not on Meta's purchase value.
- Catch fatigue early through frequency, CTR and hook rate. Meta's Creative Limited and Creative Fatigue statuses only confirm it later.
- Audit dashboards monthly. Since 12 January 2026 Meta's Ads Insights API no longer returns the 7-day view and 28-day view windows.
Meta ads optimization: frequently asked questions
How often should I optimise Meta ads?
Does changing the budget reset the learning phase?
What counts as a significant edit in Meta Ads Manager?
When should I turn off a Facebook ad?
Why did my Meta conversions drop in January 2026?
Should I include GST or VAT when calculating Meta ROAS?
Sources
- Meta for Developers: Ads Insights API metric availability updates (16 October 2025)
- Meta for Business: Simplifying ad measurement for a social-first world (3 March 2026)
- Meta Business Help Centre: About creative fatigue recommendations
- Meta Business Help Center: About significant edits and the learning phase
- PPC Land: Meta restricts attribution windows and data retention in Ads Insights API
- Good Morning Co: Meta attribution changes and engage-through
- Jon Loomer: Creative fatigue in Meta ads
- Google Ads Help: Taxes on advertising costs
Let ScaleWin read your ad data every day
ScaleWin checks your Meta and Google Ads daily, judges every ad on profit and tells you exactly what to change.
Start free, no card



