Most small businesses start paid ads the wrong way round: they pick a platform, set a daily budget that “feels safe”, and judge it after a week. A better order is numbers first, then channel, then tracking, then tests.
Step 1: know what a sale is worth
Before any campaign, write down three numbers:
- Average sale value: an order, or the first invoice from a new client.
- Gross margin: what’s left after the cost of the product or delivering the service.
- How much of that margin you’ll spend to win a sale. This is your target cost per sale.
For example, a AED 400 sale at 45% margin leaves AED 180. If you’ll spend half of it to win a customer, your target cost per sale is AED 90. Above that, ads lose money on the first sale, unless customers reliably come back.
If you sell services through enquiries, convert this into a cost per lead using your close rate. With a 20% close rate, AED 90 per customer means you can pay about AED 18 per lead.
These numbers decide almost everything else, including whether paid ads make sense at all. How much to spend on ads turns them into a testing budget.
Step 2: pick the channel by how people find you
| If… | Start with | Why |
|---|---|---|
| People already search for what you sell (“AC repair Dubai Marina”, “buy abaya online”) | Google Search ads | You meet existing demand at the moment of need |
| Your product sells on sight: fashion, food, beauty, home | Meta (Instagram, Facebook) | Visual formats create demand; strong for e-commerce |
| Your audience is younger and you can make native short videos regularly | TikTok | Works when the creative looks like TikTok, not like an ad |
| You need local calls and visits | Google Search plus your Google Business Profile | Local intent is search-led |
Start with one channel. Splitting a small budget across three means none of them gets enough data to learn.
Step 3: track the real result before spending
Platforms optimise for whatever you tell them is a conversion. If that’s a page view or a button click instead of a purchase or a qualified enquiry, you’ll get more page views and button clicks.
Before launch, confirm:
- The purchase or enquiry is recorded in Google Ads or Meta Events Manager, with the value where possible.
- Server-side tracking is in place where you can manage it: Meta’s Conversions API and Google’s enhanced conversions. See Conversions API and enhanced conversions.
- If you sell through enquiries, you have a plan to send qualified leads and sales back to the platforms. See offline conversions explained.
Step 4: give the platform enough to learn
Automated bidding learns from conversions. Two published guidelines are useful anchors:
- Meta: an ad set usually leaves the learning phase after about 50 results in the week after its last significant edit. Ad sets that can’t get there show “Learning limited”.
- Google: to evaluate Smart Bidding, Google recommends looking at periods with at least 30 conversions (50 for Target ROAS), such as a month or longer.
If your budget can’t buy 50 purchases a week at your target cost, don’t spread it over many ad sets. Use fewer ad sets, or optimise for an earlier step that happens more often (such as add to cart or a lead) while you build volume.
Step 5: test one thing at a time
A simple first-month plan:
| Weeks | Test | Keep everything else the same |
|---|---|---|
| 1–2 | Setup check: tracking firing, search terms or placements sensible, enquiries followed up | Budget, offer |
| 2–4 | Two or three creative angles, or two landing pages | Audience, budget |
| 5–8 | The winning angle with a second offer, or a budget increase of no more than about 20–30% at a time | Creative |
Meta’s guidance is to avoid unnecessary edits during learning, because significant edits reset it. Big daily changes make results impossible to read on any platform.
Step 6: decide with a stop-or-scale rule
Agree the rule before you start, so you don’t decide on mood:
- Scale when cost per sale is under target for two consecutive weeks with enough volume to trust it.
- Fix when clicks are fine but sales aren’t. That’s usually the page, the offer or follow-up, not the ad. See ads get clicks but no sales.
- Stop when cost per sale stays above break-even after the planned test budget is spent, and the fixes above have been tried.
Common small-budget mistakes
- Boosting posts instead of running a campaign with a conversion goal.
- Sending ad traffic to the homepage. A focused page usually converts better. See landing page vs homepage.
- Judging on platform numbers alone. Platforms count differently. Compare with your own orders or CRM. See why Meta, Google and GA4 never agree.
- Slow follow-up on leads. An enquiry answered the next day is worth far less than one answered within the hour.
Frequently asked questions
What’s the minimum budget for paid ads?
There isn’t a useful universal minimum. Work it out from your target cost per result and the volume the platform needs to learn, as in step 4. The starter plan does the maths.
Should we hire an agency or run ads ourselves?
If your monthly spend is small, learning the basics yourself is often worth it, as long as tracking is set up properly. Whoever runs the ads, make sure the ad accounts, pixel and data are in your business’s name. See who owns your accounts.
How long before we know if ads work?
Plan for at least four to eight weeks of steady spend, with tracking checked in week one. Judging after a few days mostly measures the learning phase.
If you’d like help setting up the numbers, tracking and first tests, that’s what our traffic sources service covers.
Free download · Excel workbook
Paid ads starter plan
Work out your break-even cost per sale, check which channel fits your business, and plan your first tests with a stop-or-scale rule for each.
Download itNo sign-up. Opens in Excel, Google Sheets or Numbers.
Rather have it done?
This is the kind of work our traffic sources service covers. Tell us where you are today and we’ll reply with a plan and a fixed price.
Let’s talk