Learn · Technology and vendors

Build vs buy for beginners

For most first loyalty programmes, buy a configurable platform. Build in-house only when you need unique control over member data and have at least two dedicated engineers, because building delays launch, adds ongoing maintenance and distracts from marketing the programme.

How it works

Building a loyalty programme means your own engineers create the member database, points ledger, rules engine, member portal and reporting tools. You pay for servers, security patches and future changes. Buying means you subscribe to a loyalty platform, configure earn and burn rules, tiers and expiry settings, and let the supplier run the software. The real difference is not the initial price. It is who carries the ongoing responsibility for uptime, data security and rule updates. For a first programme, that ongoing burden is usually greater than the upfront saving from building. A bought platform also gets you to launch in weeks or a few months, while a custom build often takes many months before the first member can enrol.

Worked example

This example uses illustrative figures only. Suppose you expect 5,000 members with an average of 2 transactions per member per month, or 120,000 transactions per year. A bought platform charges 600 per month plus 0.05 per transaction. Annual cost equals 7,200 in base fees plus 6,000 in transaction fees, total 13,200. Over three years that is 39,600. Building in-house costs 90,000 upfront plus 18,000 per year for maintenance, so over three years build costs 144,000. If the build delays launch by six months and each member would generate 2.00 in additional margin per month, the delayed launch forgoes 5,000 times 2.00 times 6, which equals 60,000. Add that to the build cost and the total three year cost of building is 204,000.

What to do

Buy a configurable loyalty platform for your first programme. Write down your must-have rules before you compare options: how members earn points, what they can redeem, how many tiers you need and when points expire. Ask for a platform that lets you export raw member and transaction data, so you are not trapped. Set a launch date and work backwards. Only consider building if you have a unique integration or data model that no platform supports, and you already employ at least two engineers who can maintain the system after launch. Do not treat a custom build as a way to learn software development.

Common mistakes

  • Building a custom programme before you know your earn and burn rules, which locks in wrong assumptions for months.
  • Comparing only setup fees and ignoring monthly platform fees plus the cost of your engineers' time over three years.
  • Choosing to build because you assume ownership is cheaper, without estimating maintenance for security patches and rule changes.
  • Assuming a bought platform cannot be customised, when you can ask suppliers about custom member fields and event triggers.

Related