Your Facebook budget should follow your allowable cost per lead — not a round number a competitor once mentioned at a networking breakfast.

The right one is: what can a qualified lead actually cost your business and still leave a profit worth having? Every "spend at least RX a day" rule of thumb skips that question entirely, which means it ignores your margin, your sales cycle and your delivery capacity, and it can just as easily push you into overspending on a low-margin product, or underspending on a high-value service that could comfortably sustain a much larger allowable cost per lead.

What exists instead of a magic number is a correct method. Work out what a qualified lead is actually worth to your business, fund a test large enough to produce a real decision, and separate media spend from the cost of running the account properly. The rest of this article walks through that method step by step.

The figures in the worked example below are illustrative only, to show how the calculation works. Replace them with your own real contribution, close rate and volume numbers before setting an actual budget for your business.

Contribution is revenue minus the direct cost of actually delivering the product or service, materials, direct labour tied to fulfilment, payment gateway fees. It is not the same number as revenue, and budgeting from revenue alone tends to overstate what you can genuinely afford to spend on acquisition.

If you sell several products or service tiers, use a blended figure based on what you actually expect to sell, not your single highest-margin item. An optimistic assumption here compounds into an unrealistic budget by the time you reach the later steps.

Once you know contribution per customer, apply your lead-to-customer close rate to work out how much a qualified lead can cost while still leaving acceptable profit and risk margin. If contribution is R4,000 and one in five qualified leads becomes a customer, the raw allowable cost sits at R800, before leaving any room for profit, risk, or the cost of actually running the account.

Leave real margin below that raw number. Treating the theoretical maximum as your actual target budget removes any room for a slower month, an underperforming ad set, or a lead that simply takes longer to close than your average.

A budget too small to generate enough leads to judge quality wastes the spend without producing a useful answer either way. Estimate how many qualified leads you need to see (often somewhere between fifteen and thirty, depending on your close rate and how much variance you can tolerate) before you can honestly judge whether a campaign is working.

Multiply that number by your working allowable cost per lead to get a realistic learning budget, and commit to running it long enough to reach that volume, rather than judging performance after two or three days of activity because someone got nervous.

If the learning budget this implies is larger than you can currently commit, that is useful information on its own. It may mean starting with a narrower, cheaper-to-reach audience or a smaller geographic area first, rather than launching broadly with a test that was under-funded from day one.

Media spend (what you pay Meta for delivery) and management cost (strategy, setup, creative production, optimisation and reporting) are two different budget lines. Treating them as one number leads straight to underestimating what a properly run campaign actually costs.

If you run campaigns yourself, your time is still a real cost, even without an invoice attached to it. If you use an agency or freelancer, get the fee structure in writing before committing any media budget, so you know exactly what that fee actually includes.

Nexus manages paid media from R5,000 per month, plus 10% of ad spend, plus a R1,800 once-off setup fee for account structure, tracking and initial creative. We disclose this here specifically so you can compare the real cost of managed campaigns against running them yourself before deciding which fits your stage and capacity better.

Review weekly for operational issues (a paused campaign, a rejected ad, a landing page error) and monthly for the allocation decision itself: scale where qualified lead cost and quality are holding up, hold where results are still inconclusive, fix where the landing page or offer is the real problem, or cut where the economics simply do not work at any reasonable spend.

Judge campaigns on qualified lead cost and downstream close rate, never on cost per click or reach alone. A cheap click that never becomes a qualified conversation is not a good result, however good it looks inside the platform’s own dashboard.

Give a genuinely new campaign or audience enough time and budget to exit the platform’s learning phase before you judge it, but set that time limit in advance, so "give it more time" does not quietly turn into an unlimited excuse to keep spending on something that was never going to work.

• Contribution per customer is based on real numbers, not optimistic assumptions
• An allowable cost per qualified lead has been calculated, with margin for risk and profit
• The learning budget is large enough to produce a real decision, not just a few days of spend
• Media spend and management cost are budgeted as two separate line items
• A weekly and monthly review cadence exists with defined scale, hold, fix or cut rules
• Creative budget exists, media spend alone cannot fix a weak or fatigued ad

The cheapest click is not the cheapest customer. Budget for the qualified lead you actually need, not the cost per click that looks good in a screenshot.

Nexus paid media principle

Spreading a modest budget across too many objectives, audiences and formats at once is one of the most common mistakes we see, it prevents any single test from reaching the volume needed to produce a real decision either way.

Judging performance purely on cost per lead, without tracking what happens after the lead arrives, is another. A campaign producing many cheap, low-quality leads can be worse for the business than one producing fewer, more expensive, genuinely higher-quality ones.

Ignoring creative budget entirely is a third. When performance flattens, the instinct is usually to raise media spend rather than refresh the offer, angle or creative that has quietly gone stale. Protect a real, ongoing budget for creative production and testing, not just for delivery.

A retail brand in November behaves nothing like the same brand in February, and treating every month against one fixed allowable cost per lead punishes you twice over, once when demand is genuinely higher and the auction gets more expensive, and again when a quiet month gets misread as a failing campaign rather than a predictable seasonal dip.

Look back over at least a year of data if you have it, even rough data, before setting monthly budget bands. A business that knows October and December run hotter than February can plan cash flow and staffing for the enquiry spike instead of being surprised by it every single year.

If you do not have a year of history yet, build in a wider margin on your allowable cost per lead for the first few months, and treat the early data as a baseline to refine rather than a verdict to defend in a board meeting.

Days 1–7: pull your real contribution, close rate and delivery capacity numbers, and calculate a working allowable cost per qualified lead with a sensible margin built in.

Days 8–21: run a properly sized learning budget against one clear audience and offer, tracking qualified leads and their eventual disposition, not just clicks or form fills.

Days 22–30: review qualified lead cost and quality against your allowable figure, decide to scale, hold, fix or cut, and set the next month’s budget and creative plan based on that evidence, not on repeating the same number by default.