How to Choose a Design Partner
Portfolios and proposals tell you about capability. They tell you almost nothing about fit. Five questions to ask before you sign, and what the way they answer reveals.
When a design engagement goes badly, the cause is usually not capability. It is that the wrong kind of partner was chosen, or that things which should have been settled at contract time were left open.
Capability you can assess from a portfolio. Fit you assess from how they answer questions.
Three options and what each is actually for
A freelancer or a small team. You talk directly to the person doing the work. Decisions are fast and the cost is lower. This works when you have someone internally who can make decisions and the scope is clear. It does not work for long engagements that need continuity, because there is no cover if that person becomes unavailable.
A studio. There is a team, and the work survives a change of personnel. Quality is more consistent and accountability is clearer. It works less well when requirements are still moving, because a defined process makes mid project changes expensive.
An agency. Strong when the problem is upstream: brand, strategy, positioning across many channels. For focused product work the people actually doing the design are usually several layers below the people you meet, and you pay for the layers.
Which one you need depends less on project size than on whether you have someone internally who can decide. If you do, a freelancer or small team works fine. If you do not, you need a partner whose scope includes the deciding.
Five questions to ask before signing
Pay attention to how they answer more than to what they answer.
"What do you think is the hardest part of this project?"
Someone who names a specific part has read your material. "Nothing in particular" or "this is all straightforward" means they have not read it, or they have not recognised the difficulty. Both are equally concerning.
"Tell me about a project that did not go well."
The point is whether they can. They do not need to name the client. Someone who can describe what went wrong and what they changed afterwards has a habit of reviewing their own work. "That has not really happened" means inexperience or evasion.
"Who is actually doing the work?"
It is completely normal that the people pitching are not the people building. The question is whether they say so unprompted. Ask to see the work of the people who will be assigned.
"What do we have to decide?"
Good partners answer this immediately. Being able to name your homework means they understand how the project will run. A partner who only says "leave it to us" is a partner who will later say "we are blocked because nothing was decided".
"After this ships, can we add a screen ourselves?"
This reveals whether operations were designed at all. If the answer is no, then every future screen is another engagement. That can be a perfectly acceptable arrangement, but you should know it is the arrangement.
What to look for in a proposal
A beautiful proposal proves they are good at proposals. That is not nothing, and it is not the thing to read for.
Read for whether it contains something you did not tell them. A proposal that restates your brief is a proposal nobody thought about. A proposal that says "this part will actually be harder than it looks, and here is why" is one somebody thought about.
Then read for what they said they would not do. Proposals that narrow the scope are generally more trustworthy than proposals that agree to everything. Agreeing to everything means nothing has been decided yet.
Three things to settle in the contract
What the deliverable is. Design files, a built product, or something in between. Ambiguity here reliably produces an argument at the end.
Who owns the files. This is frequently left unstated in Japan, and it becomes a problem the day you want another partner to continue the work.
How it ends. What happens if either side stops early. It is easiest to agree while everyone is still happy.
The honest signal
Fit matters, and it does not mean being personable.
When you describe a problem, does the partner try to understand the situation before offering a solution. That is the most reliable indicator I know of. Someone who responds with questions usually does better work than someone who responds immediately with a proposal.