Multi-tenant SaaS mistakes that show up after launch
Multi-tenant SaaS mistakes that show up after launch: isolation, roles, files, billing, and when to stop and fix.
·
3 min read
Taking software in-house after a consultancy build is a planned ending, not a jailbreak: can your team run deploys, fix production, and ship the next change without the partner holding the keys? Ownership on paper is covered elsewhere. This is about operability and handover when eng moves inside.
Bring it in-house when you have (or are hiring) people who will own the backlog, and the product is stable enough that learning the codebase is the job, not firefighting every Friday. Too early looks like one junior hire against an undocumented estate, or a “we’ll figure deploys out” plan with no staging. I’ve seen founders hire a team into a black box and call it independence. It was just a new hostage situation with payroll and a Slack channel that went quiet at 6pm.
Sometimes keeping a partner for a slice still wins: thin internal team, heavy ops, or a release that needs senior depth you have not hired yet. In-house is a destination. It is not a moral requirement for next quarter, and rushing it can cost more than the consultancy fee you were trying to escape.
Before handover day, the boring truths have to hold. Repos and cloud in your org. Infra described in git, not in someone’s memory. A pipeline your people can run. Staging that lies less than production. One-command local bring-up, or a written path close to it. Docs beside the code: how to deploy, how to roll back, where secrets live, who gets paged, and what “done” looks like for a release. If those are missing, you are not taking software in-house. You are inheriting a mystery with a burn rate.
Eight Mile’s SaaS work is built so leaving is normal: you own code and accounts from day one, handover documents and a walkthrough, we would rather be recommended than depended on, and the walkthrough is part of the job, not an unpaid favour at the end. Soft path when the slice is eng capacity inside your team: software development.
Not a zip file and silence. A real handover is time with your new engineers: architecture walkthrough, deploy together, first small change through the pipeline with them at the keyboard. Runbooks for the failures you already know. Named contacts for a short warranty window after. If the partner cannot sit with your team for a day, they did not design for in-house; they designed for renewals.
Stay if hiring would leave you thinner on-call than you are today, or if the next six months are mostly product discovery and you need continuity more than headcount. Switch when you have people ready to own the machine and the estate is already operable. Failures: agency-hosted forever; “handover” that is files in an email; hiring into a system nobody can rebuild from a clean laptop. Demand the opposite from day one if in-house is the plan, not as a surprise clause at the end of the contract.
If taking software in-house after a consultancy build is on your roadmap, demand handover-ready delivery from day one, or book a handover plan when your team is hired.
Own the accounts. Run the pipeline. Walk through with your people until they ship one change without you in the room.
If you want a build or handover structured that way, contact Eight Mile and say when you want eng inside and what still only the partner can deploy today.
Multi-tenant SaaS mistakes that show up after launch: isolation, roles, files, billing, and when to stop and fix.
·
3 min read
You don’t need a 40-page spec to start: what to bring instead, ruthless v1 must-haves, and how a first release gets priced.
·
3 min read
What “first working link in week two” actually means: a real URL for feedback, not a finished SaaS on day 14.
·
3 min read
Rescue a SaaS someone else built: when to take over an inherited product, stabilise vs rewrite, and own the accounts.
·
3 min read
Internal tools: buy SaaS or build on your data: when shelf seats win, and when custom on your stack is the real product.
·
3 min read
Customer portal vs email-and-Excel for trades / clinics / field teams: when a login beats chaser calls and wrong quote versions.
·
3 min read
When spreadsheets stop being enough for ops: breaking signs, what a proper system is, and how to replace the sheet without a scary IT project.
·
3 min read
Who owns the code when you hire a London software consultancy? You should: repos, cloud, domain and data from day one.
·
3 min read
Fixed-price software vs agency retainers: cost certainty and a finish line versus monthly capacity, and which fits your build.
·
3 min read
Do I need a technical co-founder or a development partner? Equity leadership vs a paid partner who ships and leaves you the IP.
·
3 min read
Build a SaaS in 6 weeks: when a focused MVP is realistic, what you need before week one, and how Eight Mile ships it.
·
3 min read
Field team software fails without offline-first: vans need capture, queue, and sync that works when signal dies.
·
3 min read
Security audit checklist before you go live: access, secrets, backups, auth, deps, alerting, and when to bring in an auditor.
·
3 min read
AWS vs managed hosting for a small SaaS: when simple platforms fit, and when owned cloud, IAM and infra in git win.
·
3 min read