The pattern repeats every year: a business spends its entire annual budget on a new website, launches in October, and has nothing left to make anyone aware the site exists. Six months later the verdict is "the website does not work", when what did not work was the spending model. Below is a workable way to split a digital budget across twelve months: how much of turnover to allocate, how the split changes with business stage, three budgets in actual numbers, which recurring costs quietly eat the plan, and the rule for moving money between channels.

The two halves people keep merging

A digital budget is made of one-off investment in an asset and recurring spend on a flow of customers. The first builds the tool; the second makes it work. Trouble starts when they are counted as one number: a $4,000 website looks like "this year’s marketing budget", when it is a capital cost after which marketing has not begun yet.

• One-off investment — website, identity, photography, analytics setup, automation. It lasts two to four years; reference prices for the site are in our breakdown of development costs.
• Recurring spend — SEO and content, ad budgets, site maintenance, software subscriptions. Stop paying and the flow stops with it.
• Reserve — 10–20% for testing new channels and for whatever you underestimated. Without it, one unplanned expense breaks the plan.
• How much in total: for a small business with steady demand, 5–10% of annual turnover is the benchmark; launching a new line makes 15–20% realistic in year one, because that product has not earned anything yet.

The split depends on stage, not on taste

There is no universal ratio, but there are four common situations, and in each of them money should flow in a different direction.

The logic is straightforward. At the start you are paying for somewhere to send people at all, so the asset takes half. After that the centre of gravity moves towards channels and the site line shrinks to maintenance plus targeted improvements. The content-heavy B2B row is not an accident: those buyers decide over months and read rather than click. How to divide money between organic and paid is covered in our comparison of SEO and paid search.

Three budgets in real numbers

Percentages mean nothing until they are converted into sums. The three scenarios below are what genuinely fits into that money.

1. 1

   $500 a month — $6,000 a year

   The "one person owns everything" level. Around $2,500 once for a decent site with a proper technical base, $1,200 a year for advertising in a single tightly targeted channel, $1,500 for copy and basic SEO, the rest on hosting, domains and tools. Spreading this across five channels is pointless.

2. 2

   $1,500 a month — $18,000 a year

   Two channels can now run in parallel. Roughly $5,000 for a site or a serious redesign, $6,000 on ads, $4,500 on continuous SEO and content, $2,500 for support, subscriptions and reserve. Without working analytics, half of that is spent blind.

3. 3

   $4,000+ a month — from $50,000 a year

   You can afford what is a luxury on smaller budgets: a real content plan, automated lead handling, continuous conversion work and genuine channel tests funded from the reserve. The main risk here is no longer money but the absence of one person who reads the numbers monthly.

The recurring costs that break the plan

Budgets usually crack not because advertising is expensive but because of a dozen small lines that never made it into the spreadsheet. Together they comfortably take 10–15% of the annual figure.

Reviewing the budget every quarter

An annual budget is not a document you approve and forget. It is a hypothesis you check four times a year. This sequence works even without a marketer on staff.

1. 1Pull spend and enquiries together per channel. If you cannot say where your last 20 customers came from, start with setting up analytics properly.
2. 2Calculate cost per enquiry per channel, then cost per closed deal: the channel with the cheapest enquiries often produces the worst ones.
3. 3Move 10–20% from the weakest channel to the strongest. Not the whole budget: sharp moves destroy the data you need next quarter.
4. 4Do not kill a channel before its cycle completes. Advertising shows its hand in 3–4 weeks, SEO in 4–6 months, content later still.
5. 5Spend the reserve or let it go. An unspent reserve at year end means you tested nothing new all year.

One more rule that prevents seasonal collapses: do not plan in equal twelfths, plan against your own demand curve. If your peaks are March and September, ad spend should rise in February and August rather than during the peak itself, when competition for impressions is already at its highest.

Frequently asked questions about digital budgets

In short

• Separate one-off investment in the asset from recurring spend on customer flow — different lines, different logic.
• The ratio follows business stage: half the budget goes to site and brand at launch, then the weight shifts to channels.
• Hold back 10–20% as a test reserve. An untouched reserve in December means a year with no hypothesis tested.
• Small recurring costs take 10–15% of the budget if they are not written down in advance.
• Review the split quarterly on cost per enquiry, but never judge a channel before its natural cycle is complete.