Build vs. buy a community platform: what happens after the weekend prototype

Somebody on your team spent a Saturday prompting Claude or Lovable, and by Sunday night there was a working community: login, profiles, threaded replies, a dark mode nobody asked for. On Monday they showed it around and someone said what everyone was thinking. We were quoted how much for this?

I build a lot of my own tooling, so I'm not going to tell you the prototype doesn't count. It does. It's also the easy part. Here's what usually comes next.

Week 2: the first requests

Customers shouldn't need another password, so you add single sign-on with your app. Your biggest accounts want private spaces. The community manager asks for email digests, because nobody returns to a forum they forget exists. Search needs to find answers, not only titles. Each one is a good prompt and an afternoon. Each one is also code someone now reviews, tests and keeps working every time your app ships.

Month 2: the parts nobody demos

The week it goes public, a spam bot finds the sign-up form, so you build a review queue and a ban button. A member posts a screenshot with another customer's details in it. Someone in Germany asks for a copy of their data, then asks you to delete it. A screen reader user can't get past the post editor. None of it is hard on its own. It just keeps arriving.

Month 4: other teams find it

Sales wants to see which prospects are active, inside HubSpot or Salesforce. Customer success wants an onboarding space for new accounts. Marketing asks why none of the good answers show up in Google or in ChatGPT. Then an enterprise customer's security questionnaire arrives with a section about the community, and it lands with whoever built it. That's good news, by the way. It means the community is working. It also means it's a real product with real stakeholders.

Month 9: the builder goes back to the roadmap

Nobody does anything wrong. The person who built it is good, so they get moved onto the product work that pays the bills. Dependency updates stack up. The community manager's requests sit behind real product priorities. The community still works. It just stops getting better, and members notice before anyone else does.

Read the full nine months and run the three-year math (a work email opens it)

When building is the right call

Sometimes it is. If the community is your product, like a social app or a marketplace where members talking to each other is what customers pay for, own that code. Same if you need a data model no platform supports, or you have a platform team with the time and a mandate to run it for years.

If the community supports your product by answering questions, onboarding customers and turning happy users into references, the weekend still paid off. You know exactly what you want now. That prototype is a great spec.

The three-year math

These are the calculator's starting values, and every one of them is a placeholder for yours:

• Two engineers for six months to reach a first launch: $180,000 at a $180,000 fully loaded cost per engineer.
• One engineer to run it for the remaining 30 months: $450,000.
• Hosting, email and monitoring at $1,500 a month after launch: $45,000.
• AI coding tools and tokens at $500 a month: $18,000.
• Security and compliance, including pen tests and adding the app to your SOC 2 scope, at $40,000 a year: $120,000.

That's $813,000 over three years. Bettermode Growth is $18,000 a year at list, so $54,000, plus about 80 engineering hours to set up SSO, the CRM connection and your domain: $60,923. Most of the gap is the engineer who keeps running it after launch, which is the cost the weekend never shows you. Community management is left out of both sides, because someone has to run the community whichever way you go.

If you're still comparing platforms, community website builders: how to choose one compares five, and the security page covers what a questionnaire will ask about.

Four questions to answer together

With whoever built the prototype and whoever owns the budget:

• Who looks after the community once its builder moves on to product work?
• Who answers the community section of your next customer's security questionnaire?
• Which product work would year two of upkeep replace?
• How will sales and customer success see community activity in your CRM?

Open the build vs. buy calculator and put your own numbers on the timeline

FAQs

Should I build my own community platform or buy one?

Build if the community is your product, if you need a data model no platform supports, or if you have a platform team with the time and a mandate to run it for years. Buy if the community supports your product by answering questions, onboarding customers and turning happy users into references.

Can I keep the prototype I built?

Keep it as the spec. It shows exactly which spaces, permissions and flows you want, which makes setting up a platform faster. The code itself is what turns into the upkeep.

How much does it cost to build a community platform?

With two engineers for six months, one engineer to run it after launch, hosting, AI tools and security work, three years comes to about $813,000 at a $180,000 fully loaded cost per engineer. Your numbers will differ, so run them in the calculator.

TL;DR

• A weekend prototype is real progress and a great spec. The upkeep is what it doesn't show.
• Single sign-on, moderation, data requests, CRM sync and security reviews arrive within the first few months.
• Build when the community is the product or you have a platform team with a long mandate.
• At the calculator's defaults, building costs about $813,000 over three years against $60,923 on Bettermode Growth.
• Most of that gap is the engineer who keeps it running after launch.