Does a startup need an IT strategy?
Yes, but a short one. For a startup, an IT strategy is a few pages that say who owns the accounts and the data, what you build and what you buy, how you keep it secure, and what you will spend on it.
Cloud users estimate that 29% of their cloud spend is wasted, the first increase in five years. The survey covered more than 750 cloud decision-makers and users.
From my work
At one client the team noticed a spike in the AWS bill. In CloudWatch I found queries hitting an old logs table of more than 24 GB. Most queries asked for a time range, so I partitioned the table by quarter. I also found EC2 instances nobody used any more and deleted them.
At Mom Software the architecture changes were about security basics: dropping third-party libraries that were not safe, removing hard-coded keys, and moving to an Angular version that still gets security updates.
What an IT strategy covers at a startup
- Ownership: the domain, the cloud accounts, the code repos and the app store accounts are in the company’s name, not a freelancer’s.
- Access: who can reach production, how logins are protected, and how access is removed when someone leaves, on the day and not when somebody remembers.
- Build or buy: which parts are your product and worth building, and which are commodity tools you rent.
- Security basics: secrets out of the code, dependencies kept up to date, backups that have been restored at least once.
- Spend: a monthly look at the cloud bill and the software subscriptions, including the ones somebody signed up for to try once.
What to skip for now
Enterprise frameworks, a 40-page policy set and tools built for companies with an IT department. Nobody reads the 40 pages, often including the person who wrote them. Write down the decisions you have made and why, keep it short, and review it when something changes, like a new hire or a round.
The most common mistake
Relying on third parties for almost everything. It starts with a no-code MVP, which is fine for testing an idea, and goes on long after the idea is proven. By the time it hurts, leaving takes months. So I want an IT strategy to name the tools you use and also the point where you move off each one.
Cloud cost belongs in the plan
The waste figure above is an estimate from the people paying the bills, not an audit, but it matches what I see: old resources nobody switched off, and data stored or queried in ways that cost more than they need to. A server you forgot about still sends its invoice on time. A monthly look at the bill catches most of it.
Getting one written
My 14-day Trial is a good start, for $995, one time: a look at the code and infrastructure, with findings ranked by risk in the first 5 business days. Then I fix what can be fixed in the 14 days, and it ends with a written plan for the rest and a walkthrough call. I sell this, so read it with that in mind. If you already have a technical cofounder with time to write it, you may not need me.