Ecommerce marketing budget: the complete guide

Different goals produce different budgets, so name the goal first. A budget aimed at growing the customer base looks nothing like one aimed at protecting profit in a difficult quarter, and a store that has not chosen will drift towards whichever is easier to report.

There are three common goals. Buy new customers, accepting slimmer immediate returns because the value comes later. Harvest existing demand as efficiently as possible, accepting slower growth. Or hold position while something else in the business is fixed, such as stock, margin or the site itself.

Each implies different targets and different channels. Buying customers means spending on reaching people who do not know you, tolerating a higher acquisition cost, and being judged on new customer count and payback. Harvesting means concentrating on searches and audiences already close to purchase, being judged on efficiency, and expecting new customer numbers to be flat. Holding means keeping the essentials running and putting the effort into the store instead.

Write the goal in a sentence with a number attached. This quarter we want four hundred new customers at no more than a stated acquisition cost, while holding blended efficiency above a stated level. That sentence is what makes the rest of the budget decidable, because every allocation question becomes does this serve the goal.

Be honest about capacity too. A budget that produces more orders than you can pack, more questions than you can answer, or more units than you have in stock creates problems that cost more than the extra revenue. Check the constraint before you check the money, because in many stores the real limit is the warehouse rather than the marketing.

And set the period. A quarter is usually the right planning unit, with monthly adjustment, because anything shorter turns into reacting and anything longer stops matching reality.

A share of revenue is the wrong starting point because it ignores what is left after costs, and what is left is the only money advertising can come from. Two stores with identical revenue can afford very different budgets.

Start with contribution per order: revenue minus the cost of goods, payment fees, packing, packaging, outbound delivery and an allowance for returns. Multiply by expected orders to get total contribution for the period. That figure is the pool from which marketing, overheads and profit all have to come.

Now decide how the pool is divided. Overheads are largely fixed and known: rent, wages, software, insurance. Profit is a decision, not a leftover, so name the amount you intend to keep. What remains is the ceiling for marketing.

That ceiling is the maximum, not the plan. The plan is the smaller of the ceiling and what you can spend productively, which is set by your acquisition cost and the demand available. Spending the whole ceiling into a channel that has run out of efficient volume converts margin into impressions.

Worked example. Say a store expects a thousand orders at sixty dollars, with twenty four dollars of contribution each, so twenty four thousand dollars of contribution. Overheads take ten thousand and the owner wants to keep four thousand. Ten thousand is left for all marketing, including fees and tools. If fees and tools take two thousand, eight thousand buys media. At an acquisition cost of twenty five dollars that is roughly three hundred and twenty new customers, which is the number the plan should be built around.

Running that calculation once a quarter changes the conversation from what should we spend to what can we afford and what will it buy, which are answerable questions.

Do the calculation with whoever handles your accounts. Contribution is where marketing and bookkeeping meet, and a figure both sides agree on is worth more than a more precise one only marketing believes.

Most stores underestimate their marketing budget because they only count media. Split the budget into two lines and the real cost becomes visible, which changes decisions about both.

The fixed line covers everything that costs the same whether or not you run ads this month: agency or freelancer fees, email and messaging platforms, analytics and feed tools, review platforms, design software, photography, and any retainer. These are the cost of having a marketing function.

The variable line is media: the money that buys impressions and clicks, which can be turned up or down within days.

The split matters for three reasons. It shows you the true cost of a customer, because fixed costs belong in acquisition cost. It shows you your floor: the amount you spend before a single visitor arrives, which is what you carry through a quiet month. And it makes cutting rational, since cutting media reduces sales quickly while cutting fixed costs reduces capability slowly and is harder to reverse.

Audit the fixed line once a quarter. Software accumulates in ecommerce stores, and a list of every recurring marketing charge with a note of what it does and who uses it usually finds several nobody has opened in months. Cancel those and the money moves to media without any decision about the overall budget.

Watch the ratio too. If fixed costs are a large share of total marketing, most of your money is being spent on the ability to advertise rather than on advertising. That can be correct while you are building something, and it is a warning sign when it persists.

And keep production costs where they belong. Creative production is a real marketing cost, and a channel that needs constant new material is more expensive than its media figure suggests.

Record who authorised each fixed item and when it renews. Marketing tools are bought quickly and cancelled slowly, and a renewal date written into the budget is the only thing that forces the question.

A budget spent entirely on what already works will stop working, because channels saturate, creative tires and costs rise. A budget spent mostly on experiments will not pay this quarter. The split between them is one of the few genuinely strategic decisions in a marketing plan.

Use three buckets. The largest carries proven activity: the campaigns, flows and channels that reliably return above your target. Most of the budget lives here and its job is to be maintained rather than reinvented.

The second carries improvement of what exists: new creative for a working channel, a landing page rebuild, feed work, expanding a campaign into a new product group. This is lower risk than a new channel and usually higher return than more spend on the same tired assets.

The third carries genuine experiments: a channel you have never used, a new market, a different offer. Keep it small enough that losing all of it is survivable and large enough that a test can reach a conclusion. An experiment underfunded to the point where it cannot produce a readable result is worse than not running it, because it produces a false answer you will believe for a year.

Decide the proportions from your situation. A store with one channel producing everything needs a larger experimental share, because concentration is the risk. A store already spread across several working channels needs less.

Give every experiment a written question, a budget, a duration and a decision rule before it starts. We will spend this much over six weeks and continue only if acquisition cost is below this figure. Without that, experiments become permanent small spends that nobody can justify or stop.

And review the buckets quarterly, promoting what worked into the proven line and closing what did not.

Keep the three buckets visible in whatever you report each month. Once they are blended into one number, the experimental money disappears into the proven line and testing stops without anyone deciding to stop it.

Channels are not interchangeable, so allocate by function rather than by past return alone. Judging every channel on the same immediate number will always concentrate money in the ones that capture demand and starve the ones that create it.

There are three jobs. Capture demand that already exists, which is search and shopping, where someone is looking for what you sell. Create demand among people who were not looking, which is social, video and creator work. And keep customers you already have, which is email, messaging and retention work.

Each job needs a different measure. Capture is judged on efficiency and on how much of the available demand you are getting. Creation is judged on new customer count and cost, and it will always look worse on a last click view because its effect appears elsewhere. Retention is judged on repeat rate and contribution per customer, and its cost is small enough that it is almost always underfunded relative to what it returns.

Fund capture first, because it is the cheapest revenue available and there is a limit to how much of it exists. Once you are getting most of the searches worth having, more money there does nothing, and that ceiling is what tells you it is time to fund demand creation.

Fund retention next, because it is usually the cheapest of the three per dollar returned and it raises what you can afford to pay for a new customer. A store that improves repeat rate can outbid its previous self.

Then put what remains into creating demand, at whatever pace your payback period and cash allow.

And size each channel against its efficient volume rather than its share. Every channel has a point past which more money buys worse results, and finding that point is more useful than arguing about percentages.

A flat monthly budget ignores the fact that a dollar buys different results in different months. Shape the year around when your customers actually buy, using your own sales history rather than a general retail calendar.

Start by plotting orders by week for the last two or three years. The pattern is usually clearer than people expect, and it is often not the one they assume. Some stores peak in the run up to a holiday, some in the quiet weeks after it, some at the change of season, and some track a school or sporting calendar.

Then decide how to spend against the curve. Spending more when demand is high is the obvious approach and it works, though competition for attention is also highest then. Spending in the weeks just before a peak, when acquisition is cheaper and lists can be built, is often the better investment, because those customers are then available to sell to at full price during the peak.

Quiet periods are for two things. Building the assets that peak periods consume, such as creative, landing pages and email lists. And testing, because a failed experiment in a quiet month costs less than a failed experiment in your biggest week.

Protect the peak budget from being spent early. Stores that run out of money in the fortnight that matters most usually did so because the first three quarters had no ceiling.

Also plan the tail. Demand does not stop the day after a peak, and the weeks that follow often contain cheap traffic and buyers who missed the rush.

And write the shape down as a monthly plan with the reasoning attached, so that a strong month does not automatically get more money simply because it was strong.

Check the shape against stock lead times before committing. A demand peak you cannot supply is not worth advertising into, and the ordering decision usually has to be made months before the marketing one.

Marketing spend leaves the account before the revenue arrives, and stock is usually paid for before that. Plenty of growing stores have been profitable on paper and short of money in practice, so the budget has to be checked against the bank as well as against the margin.

Map the timing. When is ad spend charged, when do customer payments settle, when are suppliers paid, and how long is the lead time on stock. In many stores the sequence is stock paid first, advertising second, customer money third, which means growth consumes cash before it produces any.

Then check the budget against that sequence, not against the monthly total. A plan that works across a quarter can still fail in a particular fortnight, and it is the fortnight that bounces payments.

Payback period is the number that connects marketing to cash. If customers repay their acquisition cost on the first order, growth funds itself. If repayment takes months, every new customer is a loan you are making, and the faster you grow the more you lend. That is a good business and a dangerous cash position at the same time, and it needs planning rather than optimism.

Build in a reserve. A month of fixed marketing costs and stock commitments held back means a bad month is an inconvenience rather than a crisis, and it stops you cutting media at exactly the wrong moment.

And watch the second order effects of growth. More orders mean more packing, more support, more returns handling and more stock. Those costs arrive with the revenue, not after it, and a budget that funds the advertising but not the consequences creates a bad customer experience at the moment you are buying the most customers.

Go through the plan with whoever manages your cash position each quarter. Marketing plans and cash plans are usually built separately, and the gap between them only becomes obvious in the month it causes a problem.

Budget decisions made in the middle of a good week or a bad one are usually wrong. Write the rules in advance, with numbers, so the decision is already made when the situation arrives.

For increasing, state the conditions. Efficiency has stayed above target for a stated number of weeks, acquisition cost is below the ceiling, stock supports more orders, and the operation can pack them. When all four are true, increase by a set step and hold long enough to read the result. Large sudden increases disturb campaign performance and make the outcome unreadable, so a smaller regular step usually gets further.

For cutting, decide the order in advance and write it down. What goes first is the experimental bucket, then media in channels performing below target, then production and tools, and last the activity that maintains existing customers, because that is the cheapest revenue you have and it takes the longest to rebuild.

Set a floor. Below a certain spend, campaigns cannot gather enough data to work and the account degrades faster than the saving is worth. Know that number so a cut does not accidentally switch the machine off.

Agree the trigger periods too. A single week is noise. Three or four weeks of a consistent trend is a signal. Writing that down prevents the pattern where budgets are cut on a bad fortnight and restored on a good one, which costs more than either decision.

And separate the two questions that get confused: can we afford this spend, and is this spend working. A profitable channel can be unaffordable this month because of cash, and an unprofitable one can be worth keeping briefly while a fix lands. Say which question you are answering.

Put the rules in writing where the person managing spend can see them daily. Rules that live in one head get applied inconsistently, and inconsistent budget changes are worse than either decision made consistently.

A budget is a forecast, and a forecast that is never compared with what happened teaches you nothing. Set a monthly review with the same short agenda every time.

Start with the four numbers that justified the plan: contribution, new customers, acquisition cost and blended efficiency. Compare them against what you expected, not just against last month. Where they differ, decide whether the cause was the plan, the execution or something outside both, such as stock or a supplier problem.

Then look at the buckets. Is the proven line still performing, has the improvement work landed, and have the experiments reached the point where their decision rule applies. Close what has failed rather than letting it continue at a small level, because small permanent failures are how a budget silently fills up.

Check the fixed line for anything added during the month. Subscriptions and tools tend to arrive between reviews and never leave.

Then reforecast the rest of the quarter with what you now know. This is the part most stores skip, and it is the difference between a budget that guides decisions and a document written in January and ignored.

Keep the whole thing on one page and keep the history. A year of monthly pages shows the pattern of your own forecasting errors, which is the fastest way to get better at it. Most stores discover they are consistently optimistic about new channels and consistently pessimistic about retention.

And revisit the goal from the first chapter each quarter. Buying customers, harvesting demand and holding position are all valid, and the right one changes with stock, cash and what the business needs next. A budget that never changes its purpose is not a plan, it is a habit.

Keep the review short and the same shape every time. A monthly meeting that changes format each month becomes a discussion about the format, and the forecast never gets compared with what actually happened.