People delay contacting an agency because they think they need a specification first. You do not. A good agency should be able to turn a problem into a scope - that is a large part of what you are paying for. What you do need is to be clear about the problem, and honest about the constraints.
Write down the problem, not the solution
The most common brief we get is "we need a new website". That is a solution. It might not be the right one.
Better: "our phone stopped ringing in March and we do not know why". Or "we get enquiries but they are the wrong kind of work". Or "our competitor appears above us for every search that matters and we cannot see why".
Problems like these are answerable. Sometimes the answer is a new site. Sometimes it is a tracking fix, or three pages of content, or a change to who answers the phone. An agency that only ever recommends the expensive option is worth noticing.
The five things worth writing down
- What is happening now - the actual symptom, with a rough timeframe. "Since March" is more useful than "recently".
- What good looks like - "twenty enquiries a month instead of six" beats "a modern, professional website".
- What you have already - the current site, who built it, whether you can access it, what you can edit.
- A budget range - yes, really. More on this below.
- Any hard deadline - a trade show, a lease ending, a season. Real deadlines change the plan. Invented ones raise the price.
Five paragraphs is enough. We have quoted accurately from less.
Say the budget out loud
The advice not to reveal your budget comes from procurement, and it makes sense when you are buying identical things from competing suppliers. Websites are not identical things.
Without a range, an agency has to guess which version of the project you mean, and the guess is usually wrong in one direction or the other. With a range, you get a proposal shaped for what you can actually spend - and if it is not enough for what you asked for, you find out in the first conversation rather than the fourth.
A range is enough. "Somewhere between two and five thousand" does the job.
Hiding the budget does not get you a better price. It gets you a proposal for a different project.
What to leave to the agency
You are not expected to arrive with these, and specifying them too early usually narrows the work badly:
- The technology. WordPress, custom, headless - that follows from the requirements, not the other way round.
- The page structure. How many pages and what they are called is an output of the work.
- The design direction. References are useful. A finished layout in a slide deck is not.
- The keywords. Bring the terms your customers use. Leave the research to whoever does it daily.
Questions to ask them
- Who owns the code, hosting, domain and ad accounts when we finish? The answer should be "you", immediately.
- What is not included? A good agency answers this quickly. A vague answer becomes an invoice.
- Who actually does the work? The person in the meeting is not always the person building it.
- How will we know if this worked? If the answer is traffic, ask again.
- Can I see something you did that did not go well? The most useful question on this list. Everyone has one. How they answer tells you more than the portfolio.
Signals worth walking away from
- A quote before anyone has asked what you do
- Guaranteed rankings, particularly guaranteed positions
- Accounts created in the agency's name "for convenience"
- Reporting that arrives as a PDF you cannot verify
- Long contracts with no exit that leaves you anything
- Discomfort when you ask what happens if you leave
A brief you can copy
If you would rather work from a shape than a blank page, this is roughly what we would want to receive. Five short sections is plenty.
1. Who you are
What you sell, who buys it, where they are, and what makes people choose you over the alternative. Three or four sentences. If you cannot answer the last part, say so - it is a useful thing to discover early, and it changes the work.
2. What is happening
The symptom, with a timeframe and any numbers you have. "Enquiries dropped from about twenty a month to six, starting around March. We changed hosting in February." That second sentence is the kind of detail that saves days.
3. What you have tried
Including what did not work. It stops you paying for the same experiment twice, and it tells the agency how you think.
4. What success looks like
In numbers if you have them, in plain language if you do not. "Enough enquiries to keep two vans busy" is a perfectly good target and more useful than a percentage nobody can justify.
5. Constraints
Budget range, deadline if it is real, who has to approve things, and anything that is off the table. If your brother-in-law hosts the site and that cannot change, say so now rather than in week three.
Reading a proposal
Proposals are sales documents. Some things in them are commitments and some are decoration, and it helps to know which is which.
- Deliverables should be countable. "Six unique page templates" is a commitment. "A beautiful, modern website" is not.
- Timelines should include your side. If content is due from you in week two, it should say so, with what happens if it slips.
- Exclusions should be explicit. A proposal with no "not included" section has not been thought through, or is hiding the conversation.
- Assumptions should be listed. Good agencies write down what they assumed, because that is where quotes go wrong.
- Ownership should be stated. If the word "own" does not appear anywhere, ask before signing.
How to compare three proposals fairly
Do not lay them side by side and compare the totals. They are quoting different things.
- List every deliverable across all three in one column.
- Tick which proposals include each one.
- For anything missing, ask that agency what it would cost to add.
- Now compare like for like - and expect the gap to shrink considerably.
Frequently the cheapest quote is cheapest because it excludes content, tracking and training. Once those are added, the ranking often changes. It is also revealing to see how each agency responds to being asked - the good ones answer with a number the same day.
What makes a project go badly
In our experience it is rarely the technology. It is these, in order:
- Content never arrives. The single biggest cause of delay. Decide who writes it before you start, and put it in the timeline.
- Too many approvers. Feedback from five people with no arbiter produces a design nobody likes. Name one decision-maker.
- Feedback in fragments. Eleven separate emails over a fortnight is expensive. One consolidated document per round is not bureaucracy - it is what keeps fixed price fixed.
- Scope added quietly. "While you are in there, can you also..." is how projects overrun. It is fine to add things. Add them explicitly, with a price.
- Launch treated as the finish. The site going live is the start of the useful part. Plan for the three months after it.
Your responsibilities, plainly
The projects that go well have a client who supplies content when they said they would, gives consolidated feedback within the agreed window, names one decision-maker, and tells the agency when something is not working rather than waiting until the end.
None of that is difficult. All of it is more predictive of the outcome than which agency you picked.
You are allowed to not know yet
"We know something is wrong and we are not sure what" is a completely legitimate starting point, and it is how a good share of our projects begin. The first job is diagnosis, and it should be a small, paid, fixed-scope piece of work that ends with a written recommendation you own - including the option to take it elsewhere.
If that is where you are, send us the problem rather than the brief. We will tell you which service actually solves it, even when the answer is fewer services than you expected. You can see how we structure the work or start with the problem.