Founders stop seeing their own recurring costs after about a year. That inattention is one of the most reliable sources of immediate margin in a small acquisition — and unlike growth, it works in week two.
Start with the card statement, not the P&L
The P&L is what the seller decided to categorise. The card statement is what actually gets charged. Pull twelve months and go line by line. You will find:
Tools nobody has opened in a year, still billing annually.
Two products doing the same job, from a migration that never finished.
Plans sized for a traffic spike that happened in 2024.
A staging environment on production-sized infrastructure.
Duplicate subscriptions from when a contractor set up their own account.
Hosting is usually the biggest single line
Small products are frequently on infrastructure sized for a load they've never had. Check actual utilisation over 30 days before you resize — but if a database is running at 4% of capacity, it's costing several times what it needs to.
Also check the egress and storage lines specifically. They're the ones that grow silently and nobody re-reads the pricing page after signing up.
Renegotiate everything annual
An email to every vendor: "we've just acquired this business, we're reviewing tooling, what's available on an annual commitment?" A third will offer something. It costs you an hour.
What not to cut
Monitoring and error tracking. If it doesn't exist, add it — this is a cost worth increasing.
Backups. Obviously.
Anything customer-facing without measuring first. The "unused" tool may be what your best accounts use.
The support contractor, until you've done a month of support yourself and know the real load.
The order to do it in
Card statement audit in week one — it's pure profit and carries no risk. Infrastructure resizing in month two, after you've seen a full traffic cycle. Vendor renegotiation whenever a renewal comes up. Contractors last, and only once you can do the job yourself.