Most vendors will write your code and stop at the door of the environment that actually serves your customers. A Core70 dedicated team walks through it: named access, deployments and incidents handled by the engineers who wrote the code, and one full-time engineer accountable for all of it. What we may touch, how fast we respond, and how you take it back are written down before we start.
It isn't headcount. Several of our longest team engagements began as one engineer. What changes is whether the environment your customers depend on stays in your hands or moves into ours — and everything else follows from that.
“We run production” is only useful if both sides know exactly what it means. This is the default split. Anything on the right can move to the left — by name, in writing — when that's what you need.
Every Core70 engagement runs on the 70/10/20 split. On a dedicated team, the 10% is what makes accountability more than a word.
Not a salesperson. Not a project manager. Your Account Owner is a full-time member of the team — usually its lead — and is the one person answerable for changes and incidents in your production environment. Ten percent of every dollar you pay goes to them.
The 10% reaches them only once your invoice is settled. They are paid when you are served, not when you are sold to — which is why the person answering for your production has every reason to keep it running for years.
If your Account Owner isn't right, you say so, and we handle it under the same procedure as any other change of personnel. Nothing in the structure makes them harder to replace than anyone else on the team.
Vendors who run a client's production sometimes become hard to leave. We wrote the following into the agreement so that never happens here. At any time, without giving a reason, and without affecting the rest of the engagement:
Ask for the current access map, the change record and our own production access log. You get them.
Within five business days we confirm in writing who holds what — and that anyone who has left holds nothing.
Ask for an inventory of any production data held outside your environment, and have it returned or destroyed.
Take any system off our list, or reassign it to yourself or someone you choose. The rest of the engagement carries on.
Appoint your own CTO or lead as the owner of production, or as an approver of restricted changes. We work for them the way we work for you.
Give notice, and we hand production over in the order you specify: inventory in five business days, structured handover, credentials returned on the date. No fee, no minimum term, no argument.
None of this is exotic. It is what a well-run internal team does — written down, so that you don't have to take it on trust.
Only the people named in the contract can reach production. Each holds an individual account you issue, with multi-factor authentication and the least privilege their role needs. No shared logins. You keep the authoritative logs; we keep our own. When someone leaves the team, their access ends before their handover does.
Routine — releases you've accepted in staging, applied in agreed windows through the pipeline, always with a rollback path. Restricted — schema migrations, infrastructure, security configuration, anything that costs you money: your written approval first. Emergency — to restore service during an outage, then reported and recorded. You can declare a change freeze at any time.
Targets during business hours. Out-of-hours standby, with its own targets, is added only where you need it. A written incident report follows every critical or major incident within two business days.
Where production holds personal data, a data processing agreement is in place before any access is granted, and it records where our engineers work from. How we handle your data →
The most expensive mistake in software services is the one the client cannot see: turnover. Keeping the same team on a mission-critical system for eighteen years — and counting — is not a cost-saving strategy. It is the entire strategy.
We don't sell teams. We start with a first project or a first engineer, and the team grows when the work does. Some of our team engagements have moved between five and twenty engineers with the seasons and back down again. The production responsibility doesn't move with the headcount; it stays with the Account Owner throughout.
If you'd like us to read your code before anything else, the free code health check is where most team engagements begin: a senior engineer reads the codebase and tells you, in writing, what's solid, what's risky and what it would take to run.
If you have a technical lead who wants to keep the keys, you don't need a dedicated team — you need a dedicated developer, and we'll say so.
If what you need is an availability percentage with service credits attached, we're not the right fit. We commit to named people, written procedures and response targets — not to a number we'd have to price defensively.
And if your data can't be reached from outside its jurisdiction at all, tell us first. Sometimes the answer is in-jurisdiction virtual desktops. Sometimes the honest answer is no.
One business day. A plain answer on scope, people and price.