Build it in-house or hand it to a partner: the honest split for a data platform

Not all data-platform work should be outsourced, and not all of it should be kept in-house. Here is the line we would draw, and why.

The question we get asked most often before a platform build is the right one to ask: how much of this should we do ourselves, and how much should we hand over. The lazy answers are both wrong. Outsource everything and you end up dependent, unable to change your own estate without raising a purchase order. Keep everything in-house and a small team spends two years learning lessons a partner already knows, on your time and your budget. The useful answer is a split, and it falls along a fairly clear line.

Where to draw the line

• Hand over the work that is hard, occasional, and where mistakes are expensive. The initial architecture. The migration. The security model. The medallion structure that everything else sits on. This is work you do rarely, where the cost of getting it wrong is high and lasting, and where someone who has done it twenty times will do it faster and safer than someone doing it for the first time. Paying a partner to get the foundations right is cheaper than paying your own team to learn on them.
• Keep the work that is frequent, close to the business, and where context is everything. Building the next report. Adjusting a measure because the business changed how it counts something. Answering why is this number different this week. This is work that happens constantly, needs to know your business intimately, and is far better owned inside than routed through an external ticket queue every time. If every small change needs a supplier, your analytics moves at the speed of your supplier’s calendar, and that is nobody’s idea of self-service.
• Share the work that needs a specialist occasionally but not always. Performance tuning, a thorny data-quality problem, a new source that needs plumbing in. You do not need a full-time expert for these, but you do need access to one when they land. This is where an ongoing arrangement earns its keep: the foundations are yours to run day to day, with a partner on call for the work that genuinely needs depth.

The mistake to avoid is the false economy at either end. A business that keeps the hard foundational work in-house to save money usually pays more, later, unpicking it. A business that outsources the daily close-to-the-business work loses the very responsiveness that made analytics worth having.

The right shape is a platform you own and can run, built on foundations someone experienced got right, with expert help available for the occasional deep problem rather than for everything. That keeps you fast, safe and independent, which is the point.

If you are deciding what to build and what to hand over, that split is worth getting right before you commit either way. Talk to us at hello@hoptonanalytics.com.