How do I choose an ecommerce marketing agency?

The most common hiring mistake is starting with the channel instead of the number. A store decides it needs help with ads, hires accordingly, and discovers a year later that the ads were never the constraint.

So work out which of the four numbers behind revenue is weakest first: visitors, conversion rate, average order value, or repeat purchases. Traffic problems and conversion problems need different work, and buying more visitors when the store loses them at checkout means paying twice for the same leak.

Write the problem down in a sentence with a number attached, then use it as the brief. Mobile conversion has fallen and we do not know why. We depend on one channel and it is getting more expensive. Most customers buy once. Our product feed keeps getting disapproved. Each of those points at a different kind of help.

Then decide the scope honestly. Some stores need one channel run well. Some need the whole picture looked at because nobody has stepped back in two years. Some need a short piece of work and then to be left alone. Hiring an ongoing relationship for a one off fix wastes money in both directions.

This is also the reason we write a free plan before quoting anything. Until somebody has looked at your store and said which lever is weakest, neither side can say what the work should be, and a price quoted before that is a guess dressed up as a proposal.

This question sorts out more than any case study, and it should be asked before anything else commercial.

The accounts should be created in your business name and stay that way: the ad accounts, the analytics property, the merchant account, the email platform, the domain and the store itself. You grant access to whoever does the work, and you can remove it. If the arrangement is the other way round, then the campaign history, the audiences, the conversion data and the learning all belong to somebody else, and leaving means starting from zero.

Ask what happens at the end. Do you keep the campaigns, the flows, the creative files, the documentation. Is there a handover. Is any of it built inside a tool you cannot access afterwards.

Ask about the creative and the content too. Who owns the photographs, the video, the copy and the licences for any customer or creator content. Licences that expire when a relationship ends are a common and expensive surprise.

Then ask about data. Where does reporting live, and can you rebuild it without them. A dashboard inside a provider's own system is fine as an extra, and a problem as the only record.

Ours is simple, because it removes the argument entirely: everything is set up in your name and stays there, including the plan itself. If you stop working with us you keep what was built, and nothing has to be untangled.

Pricing structures create incentives, and the incentives matter more than the number. Ask directly, and ask what the structure would encourage in a bad month.

A fee based on a percentage of ad spend rewards spending more, which is fine while spending more is the right answer and awkward when it is not. It also means the advice to reduce budget, or to move money from ads into fixing the checkout, costs the person giving it.

A fee based on a share of revenue rewards claiming credit for orders that would have happened anyway, and it usually needs an attribution argument every month to settle what counts.

An hourly arrangement rewards hours, and a project fee rewards finishing, which suits defined pieces of work and suits ongoing operations less well.

A base fee plus growth share scoped to agreed work rewards doing that work, and it makes the advice independent of the budget. That is what we use: a base monthly fee for the scope in your plan, plus a capped growth share on revenue we add, month to month, with no percentage of ad spend and no percentage of the sales you already make, only a capped growth share on revenue we add above your baseline. Ad spend goes to the ad platform, all of it.

Whatever the structure, get the practical details in writing. What is included and what is extra. Whether tools are billed separately. What notice period applies. Whether the price changes if the budget changes. And whether you can stop without an argument, since a month to month arrangement keeps both sides honest in a way a long contract does not.

Three questions remain, and they are the ones that predict how the relationship actually feels.

Who does the work. Not who is in the meeting, but who will be inside your ad account and your store, how many other clients they handle, and whether you have one person who knows your business or a rotating queue. Ask what happens when that person is away. Small teams and large ones can both work, but you should know which one you are buying.

What happens in month one. A specific answer is a good sign: audit, tracking check, feed rebuild, first three fixes, in that order, with dates. A vague answer about strategy and onboarding usually means the first month is spent deciding what to do, which is what the plan should have already settled.

How you will both know it worked. Agree the measures before starting, tied to the fixes rather than to total revenue. If the work is on the checkout, the measure is checkout completion. If it is on the feed, it is impressions and clicks on the products that were fixed. Agree the review points too, and what would cause either side to change course.

Then ask for something written before you commit. A specific document about your store, with findings and an order of work, tells you far more about how someone thinks than any conversation. Ours is free and yours to keep, with or without us, because some stores take it to their own team and that is a reasonable outcome.

And be wary of anyone promising a result. Nobody can commit to a number they do not control.