A good paid ads pilot is a 60-day test with a fixed budget, one clear goal, pre-agreed pass/fail numbers and a written decision at the end: scale, adjust or stop. Its purpose is not to make a profit in the first two months; it is to find out whether paid advertising can work for your business at an acceptable cost, before you commit a larger budget or a long contract.
Many small businesses either spend too little for too short a time, or spend a lot with no definition of success. This guide shows how to design a pilot that gives an answer, whether you run it yourself or with an agency.
What is a paid ads pilot for?
A pilot answers three questions. Can we reach people who want what we sell? Can we turn them into enquiries or sales at a cost we can afford? And what would it take to do more of it? It does not answer whether ads will work forever, but it reduces the risk of a larger decision.
It also protects the relationship with a provider. If you hire an agency, a pilot with agreed measures is a fairer start than a twelve-month contract. Our post on vetting a paid media agency before you sign suggests questions to ask, and a pilot is a practical way to check the answers.
Step 1: Write the pass/fail numbers before you start
This is the step most pilots skip. Before spending, write down:
- The goal. Leads, sales or bookings. Choose one.
- The maximum acceptable cost per result. Work this out from your margins. Our guide to break-even ROAS and the one on setting a cost-per-lead target both show how.
- The minimum volume. How many results do you need to trust the answer? Ten leads is thin; thirty is better.
- The decision rule. For example: “If cost per qualified lead is under €X after 60 days, we scale. If between €X and €1.5X, we adjust. If above, we stop.”
Writing it down stops you from moving the goalposts later, which is the most common way pilots turn into costly habits.
Step 2: Set a budget that can produce an answer
A pilot budget should be large enough to generate a meaningful number of results. A simple way to estimate: take your maximum acceptable cost per result, multiply by the number of results you want (say 30), and add a margin for learning. If a lead can cost at most €40 and you want 30 leads, you need around €1,200 in spend, plus perhaps 20 to 30% extra for the learning phase and testing. That means about €1,500 over 60 days, or roughly €25 a day.
If that number is more than you can spend, the pilot may need a narrower scope, a cheaper conversion or a different channel. Our post on the minimum budget for Google Ads or Meta Ads explains why too-small budgets rarely give clear answers.
Step 3: Pick one channel and one offer
Resist testing everything at once. Choose the channel with the clearest intent for your business: usually Google Search for services people actively look for, or Meta for products that benefit from visual discovery. If you are unsure, see our comparison for B2B lead generation on a small budget.
Pick one offer. A pilot is not the time to test five propositions. If the offer is unclear, work on it first; our guide to productising a service shows how to package something a buyer can understand.
Step 4: Prepare the foundations
Before launching, check:
- Tracking. Conversions recorded correctly, tested with real submissions. If you operate in the EU, review Consent Mode v2.
- Ownership. You own the ad accounts and give any agency access, not the other way round. See who should own your ad account.
- A landing page. One page for the offer, with one action.
- A follow-up process. Who answers leads and within what time?
- A way to judge lead quality. A simple column in a spreadsheet: good fit, maybe, not a fit.
Step 5: Run it in three phases
Days 1 to 14: Launch and learn
Launch with a modest set of ads and keywords or audiences. Make only safety changes: fix tracking errors, add negative keywords and correct disapprovals. Avoid frequent tinkering. Our post on how long paid ads take to work explains why early results are noisy.
Days 15 to 40: Optimise
Now make considered changes. Cut wasted search terms, pause ads that clearly underperform, test a second message and adjust the landing page if people leave quickly. Make one significant change at a time and write it down.
Days 41 to 60: Confirm
Stop making large changes. Let the account run and gather a clean read on cost per result and lead quality. This is the period that tells you what a steady state looks like.
Step 6: Review weekly, decide at day 60
Keep a one-page log with the week, spend, clicks, leads, qualified leads, cost per qualified lead and notes on changes. Review it weekly for problems, not for verdicts. If something breaks, fix it; if results are mediocre, wait.
At day 60, compare actual results to your written rule. Then choose:
- Scale. The cost per qualified lead is within target and lead quality is acceptable. Raise the budget gradually, by about 20% at a time, and keep measuring.
- Adjust. Results are close but not there. Identify the biggest constraint (traffic quality, landing page, offer, follow-up) and run a second, shorter test on that.
- Stop. Results are far from target despite sensible changes. Stopping is a success if it saved you a larger loss. Put the money into something else.
If you have clicks but no leads, use our seven-day diagnostic before deciding it is the channel’s fault.
What if you hire an agency for the pilot?
Ask for a pilot structure in writing: the budget, the duration, the metrics, the reporting schedule and what happens at the end. Check what they charge; our post on paid media agency pricing models explains the common structures. Agree that you own the accounts, can see the data at any time and can end the pilot without penalty if the terms are not met. A confident provider will agree to this. A reluctant one is telling you something. Then judge reports against our checklist in seven numbers to demand from an ad agency report.
Common pilot mistakes
- Too short. Two weeks is not enough to see anything beyond noise.
- Too many variables. Changing offer, page and channel together means you cannot learn which mattered.
- Judging by cost per click. A cheap click that never enquires is expensive.
- No lead quality check. Cheap junk leads look like success on a dashboard.
- Ignoring the sales process. If leads are not followed up, the pilot will fail for reasons that are not about advertising.
The short version
Define success in numbers before you spend. Budget enough for an answer. Test one channel with one offer, make the foundations solid, change things carefully, and decide at day 60 using your own rule. A pilot that ends with a clear “scale, adjust or stop” is worth its cost even when the answer is stop.