You can’t rationalize a portfolio you can’t see

A CIO is handed a number: take fifteen percent out of the software budget by next year. It sounds like an afternoon’s work — pull the list, find the fat, cut it. Then she asks for the list, and there isn’t one. There is a system that knows the applications, a different one that holds the contracts, a finance feed that knows what cleared, and a shared drive with a spreadsheet somebody built eighteen months ago that was already wrong when they saved it. The money is real and it goes out every month. What’s missing is the one thing she was asked to produce: a view of what it’s buying.

This is the normal state of a large portfolio, not a lapse in housekeeping. Each system was bought to do its own job well, and none was built to be read alongside the others. The application inventory doesn’t know which contract pays for each app. The contract system doesn’t know which applications a vendor actually touches. Spend lives where spend lives. Ask one straight question across all three — what are we paying this vendor, for which applications, and when does it renew — and you are reconciling three exports by hand, which means the answer is stale by the time it is assembled and nobody trusts it enough to cut against.

The cost of not seeing isn’t only the duplicate tools and the shelfware, though there is plenty of both. It is the renewal. A contract comes up on a date no one is watching, the window to change terms or walk away passes quietly, and the organization is bound for another year of something most of its users already left. Multiply that across hundreds of agreements and the budget isn’t being decided; it is being renewed by default. This is the quiet way the systems you buy to get control end up controlling you: not through a clause, but through a deadline nobody owned.

Rationalizing a portfolio is the opposite of buying another platform that promises to fix everything. It is a discipline, and it needs one thing the slide decks skip: a single live view the people making the decisions can run the program from. Every application, who publishes it, which vendors it rides on, which contracts pay for it, and the date each one renews, in one place, current, and answerable. Then, against each contract, a decision somebody owns — compete it, renegotiate it, fold it into something already in the building, or let it lapse. Not a report read once and filed. A board has to sit in front of it, set the sequence, assign the owners, and come back every month as the deadlines arrive.

The dashboard is the easy half. The hard half is underneath it: the data has to be faithful, and it has to stay tied to where it came from. An account number that drops its leading zero, a renewal date that moved because a tool reformatted it, a contract value that arrives as text in one place and a number in another — any of these and the board is steering on figures that look clean and aren’t. Getting the data out of the systems that hold it, carrying their quirks across instead of letting the tools quietly “fix” it, is most of the work. A view is only worth running the program from when every number in it traces back to the record it came from.

I built this for a major US city. Its CIO and her governance board had to set direction for the whole application portfolio, so I did the analysis and built the system they run it from: a live map of roughly 900 applications tied to 3,145 contracts and 2,396 vendors across 57 departments, down from a baseline of 1,184 apps. Twenty-four consolidation initiatives are sequenced across three phases inside it, each with an owner and a due date; about 400 per-contract recommendations each sit against a renewal or notice deadline; it signs in with the city’s own Microsoft 365, records who changed what, and the board still runs the program in it today. A portfolio nobody could see, and a board that needed to run a real cut from one view — that is the kind of problem I like being brought in on.

Get in touch See what I build →