Build vs. buy in support tooling

Every support org eventually meets the engineer who says the quiet part out loud: "We could just build this ourselves." The renewal quote had a comma in it, the per-seat line kept climbing with every hire, and suddenly a weekend of internal tooling looks cheaper than another year of somebody else's SaaS.

Sometimes they're right. More often they've priced the easy half of the decision and skipped the half that bills forever.

The seat-price math that starts the fight

Buying looks simple until you multiply. Zendesk Suite Team is cheap enough to say yes to without a meeting.

Then you want the things people actually ask for — automation, better routing, an AI copilot — and you're on Professional, with the copilot billed as a separate line on top.

And that's before anyone logs in. HubSpot Service Hub Enterprise wants a ten-seat minimum and a one-time onboarding fee just to open the door.

Per-seat pricing has one honest property: it scales linearly with headcount, which is exactly the curve you were trying to flatten. Someone runs the arithmetic for a forty-person team, sees six figures a year, and reaches for the build option.

What the build column leaves off the whiteboard

The build-versus-buy spreadsheet almost always pits a real SaaS invoice against an imaginary two-week sprint. The invoice is honest; the sprint is a fantasy. What it leaves out is the line that never stops billing: maintenance.

Internal tools don't ship and retire — they accrue. The auth integration breaks when IT rotates SSO. The macro engine needs a rewrite when the helpdesk deprecates its old API. The reporting nobody documented becomes load-bearing, then becomes a ticket of its own. Eventually the person who built it gets promoted or leaves, taking the only working mental model with them. None of that appears in the estimate, because none of it is the fun part.

Where building actually pays off

Building pays off in a narrow band, and it's worth naming precisely. Build when the tool encodes a workflow no vendor sells — the odd escalation path, the compliance step peculiar to your industry. Build when it's a thin layer over systems you already own, not a platform you now have to operate. And build when it removes far more license cost than it adds in engineering time — a real consolidation, not a fourth dashboard.

That last test is where per-outcome pricing changes the sum. Intercom's Fin bills per resolution rather than per seat.

Pricing that tracks outcomes instead of headcount can be the cheaper shape when volume is lumpy — and it removes one of the main reasons teams start eyeing a build at all. If the vendor already charges by the thing you're trying to reduce, the internal-tool case gets weaker, not stronger.

The shelfware tax cuts both ways

The strongest argument against buying isn't the list price — it's waste. Most stacks are quietly carrying licenses nobody opens.

That figure should frighten a buyer more than any per-seat number: money already spent on tools that lost the internal adoption fight. But it's a case for discipline, not for building. A home-grown tool nobody adopts is shelfware too — you just can't cancel it, and you're still paying whoever keeps it running. The healthiest move the industry has managed is consolidation: the average company's SaaS portfolio recently shrank for the first time in over a decade, trimmed back toward the tools people actually use.

The honest default

Buy the commodity; build the differentiator. Ticketing, routing, a knowledge base, a copilot — these are commodities, and paying a vendor to keep them alive is cheaper than paying yourself to. Reserve your engineers for the thin, specific layer that is genuinely yours and that no invoice can buy.

And cost the build with the maintenance tax included. A DIY tool isn't cheaper because version one was free; it's cheaper only if the total cost of keeping it alive stays under the seat count it replaces. Most of the time it doesn't — which is why "we'll just build it" deserves the same skepticism you'd aim at any vendor's demo.

If you're pressure-testing a buy decision, cost it the way you'd cost anything else — see the real cost of a ticket — and if the pitch leans on self-service deflection, check what customers actually choose before you sign.