How to forecast a Meta campaign before you spend anything

A forecast will not tell you what will happen. It will tell you whether what you are hoping for is arithmetically possible, which turns out to be the more useful question.

Most campaigns that fail were doomed at the planning stage: the budget could never have bought enough conversions to escape the learning phase, or the target ROAS required a conversion rate nobody in the category achieves. Ten minutes of arithmetic catches both, before the money moves.

The chain

Everything follows from four inputs: budget, CPM, CTR and conversion rate. The chain is deliberately simple.

impressions = (budget / CPM) × 1000
clicks      = impressions × CTR
purchases   = clicks × conversion rate
revenue     = purchases × AOV
ROAS        = revenue / budget

Worked through with plausible numbers for a $30/day campaign over 14 days:

budget       = $420
CPM          = $12        → impressions = 35,000
CTR (link)   = 1.2%       → clicks      = 420
conv. rate   = 2.0%       → purchases   = 8.4
AOV          = $65        → revenue     = $546
                             ROAS       = 1.30

If you would rather not do the arithmetic by hand, the calculator runs the same chain.

Now the interesting part. That campaign produced 8 purchases in two weeks. It needs roughly 50 optimisation events a week to leave the learning phase, so it will never get there on purchases. And at 40% margin, break-even ROAS is 2.5, so a projected 1.30 is a projected loss.

Both of those are knowable before spending a cent, and both change what you should do — not “run it and see”, but “optimise for add-to-cart instead” or “raise the budget or do not start yet”.

Where to get the inputs

CPM is the one you should not guess from memory. Take it from your own last campaign in the same country if you have one; otherwise use a category and market benchmark and take the pessimistic end. CPM varies more by country and season than by anything you control.

CTR should come from your own creative history if you have any. With no history, 1.0% is a fair planning assumption for e-commerce; use 0.8% if you want the forecast to be honest rather than encouraging.

Conversion rate is the input people get most wrong, because they use their site-wide Shopify conversion rate. That figure includes returning customers, direct traffic and email — all of which convert far better than cold paid social. Halve your site-wide rate as a starting assumption for cold traffic, and use the real paid-social figure once you have one.

AOV should be your actual average order value from Shopify, not your headline product price. If you sell a $40 product and half your orders include a second item, your AOV is nearer $55 and the forecast changes materially.

Run three scenarios, not one

A single-point forecast invites false confidence. Run the chain three times:

Scenario	CPM	CTR	Conv. rate
Pessimistic	benchmark high	0.8%	half your estimate
Central	benchmark mid	1.0%	your estimate
Optimistic	benchmark low	1.5%	your estimate × 1.3

Then ask one question of the pessimistic case: can I survive this outcome for three weeks? If the answer is no, the budget is wrong regardless of how good the central case looks.

The optimistic case has one job: showing you the ceiling. If even the optimistic scenario does not clear your break-even ROAS, the campaign is not a campaign, it is a donation — and the problem is the price, the margin or the product, not the ads.

What a forecast is actually for

Three things, and none of them is prediction.

Sizing the budget. The forecast tells you the smallest budget that produces enough events to learn from. That single number prevents the most common and most expensive mistake in small-budget advertising.

Choosing the optimisation event. If purchases cannot reach 50 a week at any budget you can afford, the forecast says so immediately, and the answer is to optimise for add-to-cart and switch later.

Setting the decision rule in advance. Writing down “if after two weeks CPA is above $X, I stop” while you are calm is worth more than any amount of judgement applied while watching money leave.

Then check the forecast against reality

A forecast you never revisit is a horoscope. After the first two weeks, put the actuals next to the projections column by column.

The column that is furthest off tells you what you did not understand about your own business, and that is the real output of the exercise. A CPM three times your estimate means you misjudged your market. A CTR at half your estimate means you misjudged your creative. A conversion rate at a third means your product page, not your advertising, is where the next month of work belongs.

---

AdPilot’s simulator runs this chain for you across pessimistic, central and optimistic scenarios, using your market, category and price point — so you see what a given budget can and cannot buy before you commit it.