Build vs. buy: a decision framework for business software

Here's something a custom software studio probably shouldn't lead with: most businesses should buy software, not build it. Accounting, payroll, email, CRM for a standard sales motion — the off-the-shelf products are excellent, and building your own is lighting money on fire.

But 'default to buy' has failure modes, and they're quiet ones. Nobody sends you an invoice titled 'the cost of working the way your vendor decided you should.' This framework is how we help clients find the line.

Where buying quietly fails

Off-the-shelf software encodes someone else's idea of how your business works. That's fine when the workflow is a commodity. It fails in four predictable places:

• The workflow is your edge. If the way you dispatch trucks, price jobs, or serve regulars is why customers choose you, renting the same software as your competitors levels your advantage away.
• Per-seat pricing meets headcount. $89/user/month is nothing at 5 users and a $170k annual line item at 160 — often more than building the tool you actually need.
• Your data becomes their asset. Third-party ordering platforms are the cautionary tale: they kept the customer relationship and charged 28% for the privilege.
• Integration duct tape. When you're paying three people to re-key data between five SaaS tools, you've already built custom software — out of humans.

The three questions

Strip the decision to three questions. Answer honestly and the right side of the line is usually obvious:

• Is this workflow a differentiator or a commodity? Commodity → buy. Differentiator → build.
• What does the off-the-shelf option cost at 3× your current scale — in seats, in workarounds, in the hours your team spends fighting it? Price the workarounds, not just the subscription.
• Who owns the customer and the data at the end of five years? If the answer is 'the vendor,' understand you're renting your own business back.

The hybrid that usually wins

Build vs. buy is a false binary. The pattern that wins most often for mid-sized businesses: buy the commodity core, build the differentiated edge. Keep QuickBooks and Gmail; build the ordering platform, the dispatch board, the client portal — the pieces where your process is the product.

One of our restaurant clients runs exactly this shape: Stripe and Twilio underneath (bought), a commission-free ordering and loyalty platform on top (built). The bought parts cost hundreds a month. The built part recaptures $41,000 a month in delivery commissions.

A test you can run this week

List every subscription your team uses. Next to each, write the monthly cost, the hours of workarounds it generates, and whether a competitor could subscribe to the identical setup tomorrow. Anything expensive, workaround-heavy, and identical to your competitors' stack is a build candidate. Anything cheap, smooth, and undifferentiated stays bought.

If the list surfaces one or two build candidates, that's normal. That's also exactly the conversation a good development partner should be willing to have with you honestly — including telling you when the answer is 'keep buying.'