How should an ecommerce marketing budget be split?

Revenue is traffic multiplied by conversion rate, average order value and purchase frequency. Every marketing pound acts on one of those four, and the useful question is which one is currently limiting the business.

Work it out with your own figures. If your conversion rate is respectable and traffic is flat, traffic is the constraint and acquisition deserves the money. If traffic is healthy and conversion is poor, spending more on acquisition buys more visitors who will also fail to convert, and the money belongs in the site.

If both are reasonable and customers never come back, the constraint is frequency, and retention work returns more than either. If order values are low relative to your cost of serving, the constraint is basket size, and merchandising, bundling and delivery thresholds are the lever.

Most stores discover their constraint is not where their budget is. Advertising is the easiest thing to spend money on, which is why it tends to absorb the budget regardless of whether it is the binding limit.

Capturing existing demand covers search advertising, shopping, brand terms and any activity that meets someone already looking. It is the most measurable spend and the most likely to be taking credit for demand you already had.

Creating demand covers paid social, video, influencer and anything reaching people who were not looking. It is harder to measure, it works over a longer horizon, and without it the capture channels eventually run out of demand to harvest.

Retaining covers email, messaging, loyalty and the post-purchase experience. It is the cheapest revenue per pound in most stores and it is capped by how many customers you have, which is why it cannot be the whole plan.

Compounding covers work that keeps returning after you stop paying: organic search content, product page and feed quality, conversion improvements, and the measurement infrastructure everything else depends on. This is the category most often cut first and most expensive to have neglected.

Splitting by job rather than channel keeps the conversation about what the money is for, and it survives a platform changing its name.

For a store with existing demand and a working site, a reasonable starting shape puts the largest share on capture, a meaningful share on creation, a smaller share on retention because it costs less to run, and a protected share on compounding work.

Then move it based on evidence rather than on the calendar. If capture channels are limited by demand rather than by budget, which impression share will tell you, extra money there buys nothing and belongs in creation or compounding.

If retention is producing revenue at a fraction of the cost of acquisition and the flows are still basic, that is an obvious reallocation. Retention work is usually people and tooling rather than media spend, so a small budget change makes a large difference.

And if the constraint is conversion, the highest returning use of the next pound is often the site itself. A conversion improvement multiplies the value of every other pound in the plan, which is why it deserves funding before more traffic does.

Review the split quarterly. Monthly reviews respond to noise; annual reviews respond too late.

Reserve a fixed portion for exploration and treat it as spent regardless of outcome. New channels, new creative directions, new markets and new offers all need funding that is not competing with campaigns that already work, because they will lose that competition every time.

Size it so a genuine test is possible. A test budget too small to reach statistical readability or to give a channel a fair run is not exploration, it is a way of concluding that nothing new works.

Protect the compounding line as well. Organic search, content and conversion work take months to show and then keep paying, which makes them the easiest thing to cut in a difficult quarter and the most expensive to have cut when the quarter after that arrives.

And check the whole plan against one number: total revenue divided by total marketing spend. If individual channels improve while that ratio does not, the split is moving credit between reports rather than moving the business.