How we work / Spend
Technology spend observability and optimisation
Every part of your technology has a cost, including the people who run it. See all of it in one agreed picture, catch what moves, and put the money where it earns its keep.
Why the old sums no longer work
For years, technology was costed the way it was bought: a big purchase, spread evenly over five years, with a share sent to each department. That worked when technology arrived as servers and licences bought once and kept for their working life.
It does not arrive that way any more. Technology is now a constant cycle of buy, build, lease and replace, and most organisations run a mix of worlds at once, each costed differently:
- On-premise: bought up front and written down over years, plus the power, space, support contracts and people that keep it running.
- Cloud: metered by use, changing by the hour, billed every month.
- SaaS: paid per seat or per user, monthly or yearly, often bought by teams on their own cards.
- AI: paid by the seat, switched on inside software you already have, and billed per token or per request.
A purchase price spread over five years cannot tell you which team used what this month, what one customer costs to serve, or why the bill jumped.
Flexera sells cost-management software. The FinOps Foundation's survey is of its own practitioners, mostly in larger organisations.
A cost for every part, people included
The answer is to give every part of the ecosystem its own unit of cost: a cost token. A seat, a device, a gigabyte stored, a thousand AI tokens, an hour of an engineer's time. Each has an agreed rate, and every team's use is counted in those units. FinOps calls this unit economics, and the TBM standard sorts the costs so shared ones can be split fairly. For AI, the FinOps Foundation calls this method token economics, or tokenomics: metering AI use by the token and tying it to the value it creates. The same idea works across the whole estate. Read more about tokenomics.
People are part of it. A system that needs two engineers to keep it running costs far more than its licence. Leave people out and the cheapest-looking system is often the most expensive one.
Built on FinOps and these cost tokens, spend management covers the whole estate, on-premise, cloud, SaaS and AI, in one picture.
What getting it wrong costs
- 1
Wrong prices, wrong position
If you do not know what it costs to deliver a product or a service, you cannot price it well. Price too low and margin leaks out with every sale; price too high and you lose the deal. Across a year of sales, small errors in cost become large amounts of lost revenue.
- 2
Bills nobody planned
When what is bought drifts away from what is used, the bills grow quietly: idle cloud capacity, seats for people who have left, tools doing the same job, AI left running.
- 3
Departments judged on the wrong numbers
When shared costs are split by headcount or by habit rather than by use, the numbers mislead. A team that uses little can carry a large share and look like it performs poorly, while a heavy user that produces little looks profitable. Budgets, hiring and investment then follow the wrong picture.
An illustrative example. A shared platform costs 100,000 a year. Split by headcount, Sales (40 of 100 staff) pays 40,000 and the Data team (5 staff) pays 5,000. Measured by use, the Data team runs 70% of the workload and Sales 5%. Sales is carrying 35,000 a year that is not its own, and the Data team looks 65,000 cheaper than it is. Figures made up to show the idea.
See it. Catch it. Control it.
See it
Every team sees what it spends, in units it recognises, and every AI use has a cost per answer or per request. Bills from every vendor are brought into one layout, so they compare like for like.
Catch it
Alerts for unusual jumps, runaway-cost alerts on AI, and a monthly check of every cost against its agreed limit. An unwatched AI service can run up its bill very quickly when use grows or something loops; the point is to know within the hour, not at the month-end bill.
Control it
An owner for every cost, agreed limits and budget limits that your teams switch on, and a 12-month renewals calendar with notice dates, so nothing rolls over unnoticed.
Optimise it
- Savings ranked by value and effort, quick wins first: changes you can make within 30 days with no contract change, each with an owner.
- Unused seats found and reclaimed; tools doing the same job spotted before the next renewal.
- Committed spend compared with pay-as-you-go, so you commit where it pays and stay flexible where it does not.
- AI costed per use and modelled at half, double and five times today's use, so scaling up is a decision, not a surprise.
Start small, prove it, expand
This is a journey, not a switch. The FinOps Foundation's advice is to start small, and grow in scale, scope, and complexity
as the value shows (FinOps Foundation maturity model, read 04/10/2026). Its Crawl, Walk, Run stages describe the path:
- 1
Crawl: one focused area
Pick one business unit, or one kind of bill such as cloud or AI. Bring its bills into one layout, agree owners and cost units, show each team its costs, and agree up front how you will know it worked.
- 2
Walk: expand what works
Add business units and bill types one at a time. Switch on alerts and limits, and hold a monthly review with finance and technology together.
- 3
Run: the whole organisation
Costs, units and rates cover the whole estate, much of it automated, and they feed into pricing and budgeting. Not every area needs to reach Run; the value it brings decides.
FOCUS gives the foundation. FOCUS is the FinOps Foundation's open, standard layout for billing data, covering AI, cloud, SaaS and data centre costs; version 1.4 is current, and the major cloud providers already export in it (FOCUS, read 04/10/2026). With every bill in one layout, the cost tokens above can be counted the same way everywhere.
Governance comes first
Spend management stands on governance. The register of every system and AI use, each with an owner, is what lets a cost be traced to the team that caused it. Without it, costs end up in a shared bucket nobody answers for. Read why governance is the foundation.
Check your own setup
A free self-check you can run on your own setup in about 20 minutes: 25 plain questions, each answered yes, partly or no. Your score fills itself in, and every gap comes with what to look at first. It works for any size, from one person to a global company; questions that only matter above a certain size are marked, and you can skip them.
No sign-up and no email address. Download it and keep it.
Which service does this
The AI and technology cost review gives you the picture once, in ten working days. Cloud spend management, AI spend management and technology spend management set it up for your team to run, with an optional monthly service if you would rather we run it. Monthly reporting keeps the review's picture current. Typical prices for all of them are on our Pricing page.
We never work inside your systems. Your teams switch on what is agreed, and we measure the result from the exports you send.
Start with a conversation.
One call to understand what you spend on and what worries you. If an engagement fits, you get a written scope and price.
Book a first call