In this article (7)
The choice matters more than the stack
Companies spend weeks comparing technologies and an afternoon comparing suppliers, then lose the project to the second decision. Almost every failed custom build we have seen failed on communication, scope discipline or a team that quietly changed after the pitch. The technology was fine. Judge the partner the way you would judge a hire, because for six months that is closer to what they are. This matters most once you have settled that building beats buying for your situation.
Watch whether they push back
The most useful signal appears in the first conversation and costs nothing to read. Do they interrogate the business problem, or take your specification as given. Do they suggest a different approach, or agree with everything. A supplier who agrees with all of it is optimising for signature, not outcome, and you will meet the disagreements later as change requests. Someone who tells you in week one that a third of your requirements are not worth building is the more expensive-sounding and cheaper option.
Meet the people who will do the work
Senior engineers in the pitch and juniors on delivery is the oldest pattern in this industry. Ask directly who writes the code, how many projects they are on simultaneously, and insist on speaking to them before signing. Ask what happens if that person leaves mid-project. A partner answers specifically; a sales process answers with reassurance about bench depth and process maturity.
Ask what would make this take twice as long
This single question separates experience from optimism faster than any reference check. Someone who has delivered these projects answers immediately and concretely: unclear approval chains, dirty data, an integration with a system whose documentation is wrong. Someone who has not will reassure you. The honest answer also tells you what to prepare on your side, which is where most of the schedule is actually decided.
Talk to us about your project on WhatsAppSettle ownership in the first meeting
Who owns the code, where the repository lives, whose name is on the hosting, what you receive if the relationship ends. A good partner has clear answers ready and no objection to writing them down. Evasion here is the strongest negative signal in the whole evaluation, because it is the one area where the supplier's interest and yours genuinely diverge. The full list is worth working through as a handover specification before anything is signed.
Check they have built something like it, not everything
A portfolio spanning every sector and technology usually means a rotating cast of contractors. What matters is whether they have solved a problem structurally similar to yours: comparable integrations, comparable data volumes, comparable regulatory weight. Ask about the one that went badly and what changed afterwards. A studio with a system connecting a supply chain operation across borders has met the problems your project will meet, whatever the industry label says.
Start with something small and real
Where the commitment is large, buy the discovery phase first as a standalone piece of work with its own deliverable. You get a process map, a scoped plan and an estimate you can take elsewhere, and you learn how they communicate under a deadline before the main contract exists. Any partner confident in their work will sell this; reluctance to be evaluated on a small piece tells you what the large piece would be like. It is also the cheapest way to establish whether a custom platform is the right answer at all.
Written by
Idennex
Strategy-first agency, Istanbul