Your marketing budget should fund the constraint, not imitate a bigger company’s channel mix.

That number gets repeated in founder forums as if it settles the question, and it settles almost nothing. A high-margin consultancy with a six-month sales cycle can fund a completely different acquisition pattern from a low-margin retailer that needs profitable transactions this week. A startup with no repeatable offer yet has different priorities again. A revenue percentage can open the conversation. It should never close it.

A genuinely useful SME marketing budget answers four questions in order: what commercial outcome does the business actually need, where is the real bottleneck right now, what can the team fulfil if demand shows up tomorrow, and what evidence would change next month’s allocation? If the honest answer to the first question is "we just need more awareness," the real first job is translating that vague feeling into a target audience, a specific buying action and a way to measure whether it happened.

Treat the budget as a small portfolio, not a single lever. Some spending creates demand right now. Some improves how much of that demand actually converts. Some builds demand for later. Some is simply overhead dressed up as strategy. Sort every line item into its real category before deciding whether it has earned another rand.

No amount of traffic rescues a page that buries the offer, a form that silently fails, an untracked WhatsApp button or a sales inbox with nobody’s name attached to it. Before another rand goes into media, check mobile load speed, first-screen clarity, whether proof actually sits near the CTA, form completion rates, confirmation messaging and whether you can even see where a click came from. None of that is website polish for its own sake, it sets the ceiling on what every future click is worth.

A workable foundation also needs an actual lead process: someone named to respond, a response-time expectation, information captured at the point of contact, and a shared definition of what counts as qualified. Plenty of businesses report "marketing is generating leads" while sales insists "we are getting nothing usable", usually because no shared definition or feedback loop exists between the two teams.

Nexus platform websites start at Launch from R3,000, Business from R5,500 and Pro from R8,000, with custom from R5,000 and ecommerce from R7,500 where the scope calls for it. Eligible projects can run on disclosed 12-month plans. Treat a website as commercial infrastructure that has to earn its place in the budget, not a decorative stand-in for an actual demand plan.

A short-cycle, thin-margin business needs a tight grip on acquisition cost and conversion, because every discount, every delivery cost and every return quietly eats into what it can actually afford to pay for a customer. A high-margin professional service can often stomach a longer content or SEO payback window, provided a qualified opportunity is genuinely worth the wait. Start with unit economics: average revenue per customer, gross contribution, close rate, repeat rate and how much of that demand you can actually serve. From that, set an allowable cost of acquisition, it will be a rough guess at first, and a rough, honest guess still beats copying a competitor’s advertising cadence with no numbers behind it at all.

Cash flow imposes its own, separate ceiling. A campaign that looks profitable on a spreadsheet but needs three months of spend before any cash lands may simply not suit a business paying salaries weekly. In that situation, favour work that improves conversion on existing traffic, protects demand already in motion, and produces faster feedback than a slow-burning brand play.

Paid media is the controllable lever: fast to switch on, useful for validating an offer, and good for seasonal pushes, but it needs creative, landing pages and real follow-up to be worth anything. SEO is the compounding investment: technical health, genuinely useful service and question pages, local relevance where it applies, and conversion work. It asks for a longer runway in exchange for reducing your dependence on winning an auction every single time someone searches.

Email and CRM follow-up are chronically underfunded, mostly because their cost is less visible than a media invoice. A reliable response sequence, one genuinely useful follow-up message and a clean lead tracker can lift the return on traffic you have already paid to acquire, without spending another rand on reach. Content, meanwhile, should be doing a specific commercial job: answering the questions buyers actually ask, backing up your proof, helping sales close, and creating pages that can earn relevant search visibility over time.

Resist the urge to force every channel into the plan at once. A small team running one demand channel properly, alongside a steady conversion and content rhythm, will consistently outperform a modest budget spread thin across paid search, paid social, SEO, influencers, events and half a dozen tools nobody has time to actually learn from.

Cut on evidence, never on embarrassment. Start with subscriptions nobody actually opens, duplicated reporting tools, generic content calendars with no commercial intent behind them, campaigns that consistently produce unqualified leads, and broad targeting nobody can tie to a real outcome. A channel with genuine demand but a weak landing page or a slow sales handoff often deserves a repair before it deserves a cut.

Protect whatever preserves your ability to learn: analytics, conversion tracking, lead disposition records and a small test budget tied to a real hypothesis. Cutting measurement because results look poor just makes the next decision less informed than the last one. A Pretoria B2B firm watching its paid cost-per-lead climb should not automatically abandon the channel, it should first check search terms, qualification questions, landing-page promise and close rate, and only pull spend once those genuinely stay weak after an honest attempt to fix them.

1. Read the business constraintIs the issue insufficient demand, low landing conversion, poor lead quality, slow sales response or limited delivery capacity?
2. Review qualified pipelineLook at leads contacted, qualified, opportunities and wins by source, not only clicks or reach.
3. Inspect conversion assetsReview top landing pages, forms, WhatsApp, calls, proof, offer clarity and response time.
4. Choose scale, hold, fix or cutEvery line item should receive one decision and an owner, rather than rolling forward by default.
5. Set the next hypothesisState what will change, how much will be spent, what evidence counts and when the test stops.

Name an accountable owner for each channel, even where an outside specialist actually does the work. That owner does not need to write every ad or article personally, they need to know the objective, hold the line on trade-offs, and make sure sales feedback actually reaches the people running the campaign. A dashboard with nobody making a decision from it is just a more colourful spreadsheet.

Keep a simple monthly scorecard: spend, qualified leads, opportunities, wins, contribution (estimated or realised), conversion rate, the key learning and the next action, and resist manufacturing false precision while the data is still thin. Every quarter, remove whatever no longer earns its place. Tool sprawl is a common SME tax: a handful of subscriptions each promising automation, while the actual lead process behind them stays manual and inconsistent. New technology should only enter the stack when it removes a specific, named operational pain.

• A primary offer, audience and CTA are clear on the relevant landing page.
• Forms, call tracking and WhatsApp actions are tested and recorded.
• Sales can respond within an agreed time and record lead disposition.
• The business knows its approximate close rate, gross contribution and delivery capacity.
• Media spend is separated from agency, creative, website and tool costs.
• SEO work has a defined cluster, technical baseline and conversion goal, not a vague article quota.
• Every experiment has a cap, owner, evidence threshold and stop decision.

A budget is working when it makes the next decision easier. If it only produces activity reports, it is not yet an operating system.

Nexus growth planning principle

Days 1 to 30: define the offer precisely, fix the main conversion leaks, install tracking that actually works, and agree on what "qualified" means with whoever closes the deals. Days 31 to 60: run one focused demand initiative and improve the landing page using real responses, not guesses about what visitors probably want. Days 61 to 90: compare qualified-lead economics across whatever you tested, build the next highest-value SEO or conversion asset, and decide what has earned a bigger allocation. This produces actual evidence before channel sprawl, instead of a promise of instant scale nobody can back up.

Keep committed spend separate from variable spend in your own head, not just on paper. Committed spend covers platform licences, retained specialist capacity and essential infrastructure. Variable spend covers media, production bursts and experiments you can adjust on short notice. Get this distinction wrong and an SME can lock so much into fixed marketing overhead that it has no room left to react to a seasonal dip, a new offer or a genuinely bad month.

Write down the assumptions sitting behind each allocation (expected conversion rate, cost per click, landing conversion, sales response time) so the budget can actually be tested rather than defended as a guess dressed up in a spreadsheet. Founder hours spent chasing approvals or hand-building reports are a real cost too; if a channel constantly needs executive rescue, simplify it, name an owner, or accept that it is quietly eating time that should be going to customers.

In a downturn, evenly trimming every activity is rarely the right move. Protect whatever preserves trust and captures demand already in motion, accurate web pages, a responsive enquiry path, local listings, follow-up, and your highest-confidence demand source. Pull back speculative expansion until the evidence improves, and make every change visible and documented, so the team can actually tell what caused what instead of guessing after the fact.

Run the marketing budget calculator to sketch a working range, then adjust it against real margin, cash flow and delivery capacity. If the website or the lead process is the actual constraint, fix that before treating media spend as the answer to a problem it cannot solve.