The brief is the cheapest lever you have in this process, and almost nobody uses it. A store sends a paragraph saying it wants to grow, wants better ads, wants a fresh look, and then wonders why the three proposals that come back all look the same and none of them engages with the actual business.
They look the same because a vague brief can only be answered with a standard offer. Nobody can propose something specific about a business they know nothing about, so they propose their usual package with your name at the top.
Send something better and the responses separate immediately. You will see who read it, who did the arithmetic, and who ignored the constraint you spent a paragraph explaining. That is worth far more than the meeting where everyone is on their best behaviour.
This is what to put in, and it applies whoever you send it to, including us. None of it requires giving anyone access to your accounts.
Lead with the four numbers, even if they are rough
Monthly visitors, conversion rate, average order value and the share of customers who buy more than once. Those four multiply into your revenue, and anyone competent will ask for them anyway.
Round numbers are fine. Approximate is fine. What matters is that all four are there, because a proposal built on one of them is a proposal about one channel. If you know your conversion rate is well below what your product should command, you have already directed the conversation somewhere useful. If you know most of your customers never come back, you have directed it somewhere else entirely.
Add the traffic split while you are there: roughly what share comes from paid, from search, from email, from social and from direct. A store with most of its traffic from one paid source has a different set of risks from one with a broad mix, and the advice should differ.
If you genuinely do not have these numbers, say so, and say what you do have. Not knowing your repeat rate is itself a finding, and it tells the reader that the measurement work comes before anything else.
A brief made of adjectives gets you a proposal made of adjectives.
Say what an order is worth to you after costs
This is the part almost every brief leaves out, and it is the part that determines whether any of the advice is affordable.
Give a gross margin figure, or a range across your main product lines if they differ. Say what you pay to ship a typical order and whether the customer pays it. Say what your return rate looks like if returns are a normal part of your category. If you sell consumables, say roughly how often a customer reorders.
From those, anyone can work out what you can afford to pay to acquire a customer, which is the number every ad recommendation depends on. Without them, a proposal can only talk about return on ad spend, which is a ratio on revenue and tells you nothing about whether the revenue was worth having.
You do not have to publish your cost structure to the world. A range is enough. But a brief that says we want profitable growth without saying what profit looks like on one order is asking for a plan nobody can price.
If you would rather not share margin at the brief stage, say which of two products has more room and by roughly how much. Even that helps.
List what you have already tried, and what happened
This is the section that saves everyone the most time, and the one people leave out because it feels like admitting failure.
Write the last two years as a short history. Which channels you have run, roughly what you spent, what worked, what stopped working and when. Which apps you installed and whether they are still on. Whether you have run a redesign, a replatform or a rebrand, and what happened to revenue afterwards. Whether you have worked with an agency before and what went wrong.
The reason to include the failures is that they are evidence. A campaign that worked for a year and then stopped tells you something specific about that channel and your product. A flow that was built and never turned on tells you something about capacity in your team.
And it protects you from the most common waste in this industry, which is paying someone to rediscover what you already know. If a proposal comes back leading with an idea you tried in detail two years ago, you have learned something about how carefully it was read.
Name the constraints, they shape the plan more than the goals
Goals are easy to write and largely interchangeable. Constraints are specific to you, and they are what makes a plan realistic.
Stock and lead times: what is your reorder cycle, what sells out, what can you not promote because you cannot get it. Capacity: how many orders a day can you pack before something breaks. Seasonality: which months carry the year, and which months you would not want to be experimenting in. Team: who can change the theme, who writes copy, whether a developer is available and how quickly.
Rules: any category restrictions on advertising, claims you are not permitted to make, brand or retailer agreements that limit pricing or promotion. Technology: your platform, your email tool, anything you cannot change and anything you are already planning to change.
And the money constraint, plainly. A budget range is not a weakness to hide, it is the boundary that makes a proposal answerable. Without it, you will get a plan sized for someone else's business, and the conversation will start over once the number appears.
Say who decides and how quickly
This sounds administrative and it is one of the strongest predictors of whether the work will go well.
Name the person who can approve a change to the product page, the person who can approve a change to the budget, and the person who can approve the engagement. Sometimes that is one person, sometimes three, and the difference changes how the work has to be sequenced.
Say what your approval rhythm looks like. If every page change waits for a weekly meeting, a plan built on rapid iteration will stall. If you can act the same day, quick reversible fixes become the natural first moves.
Say when you need this to start moving and what event is driving it. A store preparing for its peak season has a different order of operations from one in its quiet quarter, and the honest answer is sometimes that a particular piece of work should wait.
Ask for reasoning back, not slides
Say in the brief what you want to receive, because it sets the standard for the response.
Ask for a written diagnosis before any recommendation: which of the four numbers is weakest and what evidence points there. Ask for the first few pieces of work in order, with the reasoning for each, and what each is expected to move. Ask what they would leave alone, which is often the most revealing question you can put to anyone looking at your store.
Ask how they charge and whether the fee changes with your spend or your sales. Ask whose name the accounts will be in. Ask what the monthly report contains and ask to see the shape of one.
Then read what comes back for whether it engages with your constraints. A response that never mentions your lead times, your margin or the season you named has not read the brief, and the working relationship will feel like that too.
We write our version of this as a four-page plan, free and yours to keep. Whoever you send your brief to, the test is the same: does the answer describe your store, or could it have been sent to anyone?
What you do not need to send
You do not need to hand over logins to be given a serious opinion. Numbers and screenshots are enough for a first diagnosis, and anyone asking for full account access before there is an agreement is asking for more than the conversation requires.
You do not need to have your data perfect. Missing numbers are informative. If nobody can tell you your repeat purchase rate, write that sentence in the brief.
One last thing. Keep the brief. Six months into any engagement it is the document that tells you whether the work went where you needed it to go, or where it was easiest to go.