The first marketing budget is the hardest one to spend well, because everything is pitched at you and nothing has a track record yet. Here's the allocation we'd defend for a typical small business — service or ecommerce — putting its first serious $1,000 a month to work, and the reasoning behind each line.
The principle first: early money should buy assets before it rents attention. An asset (a page that converts, an email list, a review base, a set of photos) keeps working after the invoice. Rented attention (ads) stops the moment you stop paying. Ads absolutely have a place — but pointed at assets, not instead of them.
Month one and two: build the machine (~$1,000 each)
- $400 — the landing page / site conversion pass. Before buying a single visitor, make the destination worth the fare: clear offer, proof, one action, working mobile layout. The full checklist is in the one-page funnel. Ads pointed at a weak page just measure the weakness faster.
- $250 — photography. One half-day of real photos of your product, your space, your face. These feed the website, ads, Google profile, and every social post for a year. Stock photos are camouflage; real ones are proof.
- $200 — email foundation. Signup form, welcome sequence, and (for stores) the abandonment flow. These flows are the closest thing marketing has to compounding interest, and they cost almost nothing to run once written.
- $150 — Google Business Profile + review engine. Complete the profile properly, then set up the habit: every happy customer gets the review link, same day. For local businesses the map pack is the highest-intent free traffic in existence.
Month three onward: the steady state
- $600 — one paid channel, done properly. One. For visual products, Meta ads with the simple structure. For "emergency + suburb" services (plumber, locksmith, repairs), Google Search ads on buying-intent terms instead. Splitting $600 across both channels gives each too little signal to optimise; concentration wins at this scale.
- $250 — content that compounds. One genuinely useful article or guide a month, aimed at a question your customers actually ask, plus the social posts cut from it (batching makes this an afternoon). Slow channel, compounding channel.
- $150 — the measurement habit. Whether it's your time or someone else's: conversions tracked, the five GA4 reports read monthly, and one number written down — cost per enquiry or per order. This line is what makes every other line accountable.
What deliberately isn't on the list
- Logo redesigns and rebrands. A consistent, adequate brand beats a beautiful inconsistent one, and at this stage brand money is conversion money in a nicer jacket. (Consistency is the cheap version.)
- Six social platforms. Pick the one your customers actually use; post-everywhere tools mostly help you be ignored in more places.
- SEO retainers before the site converts. Ranking a page that doesn't convert is renting a billboard for a shop with the lights off.
- Anything sold as "guaranteed #1 on Google." That cold email is the marketing industry's smoke alarm test — if a supplier leads with it, everything else they say is now suspect.
The two-question monthly review
At the end of each month, answer two questions in writing:
- What did an enquiry/order cost this month? (Total spend ÷ total enquiries. Crude beats absent.)
- What asset did we add? (A page, a flow, twenty reviews, a photo library — something that outlives the month.)
If the first number trends down while the second list grows, the budget is working — keep concentration and raise it when a channel proves out. If cost-per-enquiry is flat and no assets accumulated, the money leaked into activity instead of assets, and the fix is the allocation above, not more budget.
The uncomfortable summary: at $1,000 a month you can't do everything, and that constraint is the strategy. The businesses that win small budgets are the ones that do three things properly instead of nine things at a coin-toss.
Most lines on this list — the page, the flows, the profile, the content — are exactly the size of a taskblitz task. Send one over and see.