How we think

Audience first

We design for a named person, not for a committee.

Where the budget allows it, we build avatars before we build anything: who these people are, how they spend a week, what they are afraid of, what season of life they are in. For a church that means the people in that district rather than churchgoers in general. For something global it means which countries, which ages, which pains.

Then we keep asking one question for the rest of the project. John is our user: what does he do when he meets this screen, this feature, this button? It is a sobering question, and it is the only thing we know of that reliably stops a team from building what it personally finds interesting.

Traffic first

Reaching the audience costs more than building the thing. We plan for that at the start, not at the launch.

Our community has a habit of building first and asking where the people will come from afterwards, and then trusting God for the part He has not undertaken to do. We have watched good projects die of exactly this, and the budget was never the reason.

When a budget is found for development, the honest next sentence is that filling it with people will take as much again, or more. Better to hear it while there is still a decision to make.

Preparation decides

A real discovery phase saves at least a fifth of the budget, and it is the cheapest part of any project.

What it saves is the thrashing: the decisions taken twice, the screens built and then deleted, the month spent finding out in code what three meetings would have settled. Spend the meetings.

It also makes the project something a board can decide on, because at the end of it there is a scope, a design and a number instead of a feeling.

The full cycle

A project with a dead end in it loses people quietly, and nothing in the reporting says so.

We ask the whole chain out loud now. When someone fills in that form, does an email actually arrive? When a request lands, is there a person who sees it? If an ad brings somebody to a Bible school, do the first three lines hold them, and is there enough there to keep them? Every step has to have a next step.

Eighty, twenty

Eighty percent proven, twenty percent new. That ratio is what makes the dish.

Some innovation is real and we want that work: AI where it belongs, learning that is genuinely interactive, Scripture presented like something made this decade.

But innovation drifts into daydreaming, and we have had the dreamer projects to prove it. What holds is the joint of the two, with most of the weight on solutions already known to carry it and the new part placed where it earns its risk.

Design that works

Fresh, readable, unhurried. Not a portfolio piece.

We want a site to feel like it was made this year. But fresh and unreadable is a failure, and so is fresh and overloaded, and both are easy to ship because they photograph well.

Design also belongs to its audience. A design for children and a design for a conference platform have almost nothing in common, and the way to tell which is right is to ask whose eyes will be on it and what it is for.

Iterations with real users in them

Big plans are for power stations. Everything else moves in small steps.

A piece, a hypothesis, a check, an improvement. The part usually missing is the check, and the only check that counts is feedback from people who actually use the thing. A loop without them is just a schedule.

A plan for years, not for the launch

Whatever gets built has to be kept, developed and brought back into line with a market that moves.

This is partly about money, and we would rather say so. A project has a cost after launch, and a plan that stops at launch quietly hands that cost to the ministry.

So we think in years. The plan will change a hundred times; what it gives you is a bearing — where you are going, what you are trying to reach, and how you will know whether it is working.

An ugly truth beats a sweet lie

We read the metrics as they are and we tell you what they say.

If people are leaving, the useful move is to admit it and find out why, not to find a flattering way to describe the number. If the project has turned into a leaking bucket, the work is finding the holes.

That is an ordinary season in a project rather than a verdict on one. Every honest project has had it.

Room to be wrong

Everybody is wrong sometimes, including us. We build the kind of relationship where that costs nothing.

A decision that can be revisited in three months without anyone losing face usually is revisited, and the project is better for it. A decision that cannot be gets defended instead of fixed.

So: talk more, agree where you can, forgive where it is possible. These are brotherly relationships before they are commercial ones, and they outlast the project either way.

AI, used carefully

Actively, and with a clear line around what it is allowed to see.

AI did not take over the world, and it is not going to write a ministry for you. It is a very good tool, it is moving quickly, and some of it is genuinely frightening.

So use it for the Gospel, use it wisely, and be careful with data it has no business touching. That second half is the part most people skip.

Prayer over technology

However much we invent, if God does not allow it, nothing happens.

We have watched simpler, humbler solutions outlast elaborate ones more than once, and the difference was not technical. It was that the team and the partner were praying and were willing to be led.

This one is last on the list because it sits underneath the other eleven.