Guide · Ownership

Own it or rent it? What the software you already pay for really costs

The three-year arithmetic for a team of twenty, the price rises nobody budgets for, what leaving a tool actually involves, and the four cases where renting is still the right answer.

11 min readVerified
A pixel-art key.
Summary

For a team of twenty, three mainstream tools at list price cost about £73,800 over three years before a single price rise, and software prices rose 13.2% a year in the latest index, five times general inflation. A focused version one you own costs from £10,000 to build and around £800 a year to run, and nothing per seat. Renting still wins when the workflow is standard, the seat count is small and stable, or the process is still changing every month. Owning wins when the workflow is yours, the team is growing, or the data has to be in your name. The break-even question is simple to ask, and we will answer it for free on a twenty-minute call.

The short answer

Rent software when the tool fits the way you already work and the seat count is small. Own it when the workflow is yours, the team is growing, or the data has to be in your name. Most established teams are somewhere in between, paying for the workaround while the case for owning quietly builds, and the point of this guide is to put numbers on that.

The numbers are not close. A team of twenty on three mainstream tools spends about £73,800 over three years at today’s list prices, before a single price rise, and the latest index puts software price rises at 13.2% a year1.

A focused version one that you own costs from £10,000 to build and about £800 a year to run, with nothing per seat. The cost guide has the bands; this guide has the comparison.

SaaS price inflation
13.2%a year
SaaS spend per employee
$9,100a year
Hit by unexpected charges
78%of IT leaders

What renting actually costs

A subscription has four costs and the price list shows one of them. The seat price is the one on the page. The tier is the second: the feature you actually need is usually one plan up, and everyone on the tool moves up with it.

The third is usage, the AI feature or the automation task that is metered on top of the seat. The fourth is time, because the price you agreed is not the price you will pay in year three.

Take a twenty-person team on the three tools that most often turn up in a first conversation: a CRM at the professional tier, a database tool at the business tier, and an automation tool for the team.

At list prices on annual billing, HubSpot’s Sales Hub Professional is $90 a seat a month5, Airtable’s Business plan $45 a seat a month6, and Zapier’s Team plan from $69 a month7. Converted at the European Central Bank reference rate and rounded, that is the table below.

ItemYear 1Three years, flatThree years, rising 13.2% a year
CRM, 20 seats, professional tier5£16,000£48,000£54,600
Database tool, 20 seats, business tier6£8,000£24,000£27,300
Automation, team tier7£600£1,800£2,000
Rented, total£24,600£73,800£83,900

List prices in dollars on annual billing, converted at £1 = $1.358 and rounded to the nearest £100. The last column applies the 13.2% annual rise from the Vertice index to years two and three as an illustration; your renewal letters will say what actually happened.

Two things about that table. It has no usage charges in it, and 78% of IT leaders reported unexpected charges for AI features or consumption pricing in the past year2. And it is three tools. The average organisation in the same index runs 305 applications, with a median of 2402, and by the end of 2025 was spending about $9,100 a year on software for every employee1.

Twenty people at that rate is $182,000 a year, which is not a workaround any more, it is a line item.

The tier tax

The feature you need is one plan up, and the whole team moves up to get it. A single permission setting, a report or an integration can add a third to the bill for everyone. It is the most common reason a subscription that looked cheap at ten people looks expensive at twenty.

What owning actually costs

Owning has three costs, and we would rather you saw all of them than only the first.

  • The buildFrom £10,000 for a focused version one, fixed from a blueprint before you commit. Most of what established teams ask us to scope lands between £10,000 and £50,000, and the five things that move it are integrations, data, complexity, roles and compliance. The cost guide has the bands and a worked example.
  • Running itHosting, database, email and monitoring in your own accounts, at list price around $91 a month for a version one9101112, billed by the providers to you. Priced by usage, not by seat, so the twenty-first person costs nothing.
  • Keeping it workingA warranty window in which anything that does not do what the specification says is fixed at no cost, then a fixed monthly support fee if you want one. New features are scoped and priced before they are built. There is no percentage-of-build maintenance charge, and if you stop paying for support the software keeps running in your accounts with any developer you choose.

So the owned column of the comparison is a build in the first band, say £18,000, plus about £800 a year to run, which comes to roughly £20,400 over three years with no support fee, or a little more with one.

Against £73,800 rented, rising. The two lines cross somewhere in the first year, and after that the gap widens every month, because one side has a per-seat meter and the other does not.

Rented, three years, no rises£73,800
Owned, three years, run costs included£20,400
Fig. 1Three years for a team of twenty: three rented tools at list price with no price rises, against a first-band build you own plus its run costs. Support is a choice on the owned side and is not drawn.

The honest caveat is scope. A £18,000 build replaces the workflows you scoped, not every feature in three products. If the team genuinely uses a hundred features of the CRM, it should keep the CRM. The comparison holds when what the team uses is the workflow, and the rest of the tool is there because it came in the box.

The cost of leaving

Every tool has an exit cost, and it is paid whether you leave the tool or the tool leaves you, through a price rise, a discontinued plan, or an acquisition. It is worth knowing what it is before it is due.

Records come out. HubSpot, to take the largest example, exports contact records with their default and custom properties and history, plus activities and associations, delivered as a file by email4.

What does not come out as a file is everything the team built inside the tool: the automations, the views, the lists, the way work moves from one person to the next. A spreadsheet of contacts is not a CRM. The workflow is rebuilt wherever you go next, and that rebuild is the real exit cost.

The owned side has an exit cost too, and it is deliberately small. The code lives in a repository in your name from the first week, every account is in your name, and the data is yours in law under a processor agreement as well as in practice.

Leaving us means opening the repository with someone else. We think that is the right way round: the cost of leaving should sit with the supplier, not the customer.

When renting wins

Some calls end with us telling you to keep the subscription, and a guide that only argued one way would not be worth your time. Rent when:

  • The tool is the category and your workflow is standard. Accounting, payroll, email, payments. Nobody should build their own ledger. The vendor’s whole business is that workflow, and yours is something else.
  • The seat count is small and stable. Five people on a £45 seat is £2,700 a year. The maths that makes owning cheaper needs either more seats or a workflow the tool cannot do.
  • The process is still changing every month. Software fixes a process in place. Rent while you are still finding out what the process is, and build the version that survived.
  • The value is the ecosystem, not the screen. If what you are paying for is the integrations, the marketplace, the templates and the community, a bespoke build gives you the screen and none of that.

There is a fifth, quieter case: the team likes the tool. That is worth something, and we will say so.

When owning wins

Own when the workflow is yours and the tool is bending to fit it, because the bending is where the cost lives. Five signals, and they are the same five that move a build between price bands.

SignalRentOwn
SeatsFew, and steady.Growing, or the whole company plus customers and suppliers, who cannot be seats at all.
The workflowThe tool’s way is fine.Your way is the point: approvals, exceptions, the thing that makes you better than the next firm.
IntegrationsThe tool connects to what you use out of the box.Every connection is a paid add-on or a manual re-key.
DataNothing sensitive, nothing regulated.Personal, financial or health data that has to be in your name, in your region, under your policies.
ChangeMonthly.Settled enough to write down, and worth building once.

Three of five in the right-hand column and the blueprint call will almost certainly end in a firm range. Three in the left and it will end in us telling you to keep the tool.

The break-even question

What does the workaround cost over three years, in seats, tiers, tasks and hours, against a build you own? When the answer is close, keep the tool. When it is not close, it is usually not close by a wide margin, and the rises in the meantime only move it one way.

If you have already tried to close that gap with an AI builder and stalled, that is a different conversation and there is a guide for it. The arithmetic is the same; what changes is what you already have in hand.

Bring the subscriptions and the spreadsheet

Twenty minutes, no preparation, and a straight answer: keep the tools, or a firm range for the system that replaces them.

Apply for a blueprint call

Questions people ask before deciding

Sources

Every third-party figure above was read from the page cited, on the date shown. Vendors change prices without notice, so the date is part of the fact.

  1. 1Vertice, SaaS Inflation Index 2026, January 2026. SaaS prices rising 13.2% a year, nearly five times G7 inflation; 60% of vendors mask rises; about $9,100 spent on SaaS per employee by the end of 2025, from $7,900 in 2023. Seen 4 September 2026.
  2. 2Zylo, 2026 SaaS Management Index. 305 applications on average, median 240; 78% of IT leaders reported unexpected charges tied to AI features or consumption-based pricing; 61% cut projects due to unplanned SaaS cost increases. Seen 4 September 2026.
  3. 3Cledara, average SaaS spend per employee in 2026, 25 March 2026. From 1.8 million purchases across 6,800 tools; a fifty-person US company at about $6,980 a year per employee, a hundred-person company at about $3,490. Seen 4 September 2026.
  4. 4HubSpot knowledge base, export contact data. Default and custom properties with history, activities and associations, delivered as a file by email; values stored elsewhere may be redacted. Seen 4 September 2026.
  5. 5HubSpot, Sales Hub pricing. Shown in dollars: Professional $90 a seat a month on annual billing, $100 monthly; Enterprise from $150. Seen 4 September 2026.
  6. 6Airtable pricing. Team $20 and Business $45 a seat a month on annual billing. Seen 4 September 2026.
  7. 7Zapier plans and pricing. Professional from $19.99 a month, Team from $69, tasks bought in tiers from 750 a month. Seen 4 September 2026.
  8. 8European Central Bank, euro foreign exchange reference rates, 3 September 2026. USD 1.1615 and GBP 0.86055 to the euro, which makes £1 about $1.35. Seen 4 September 2026.
  9. 9Vercel pricing. Pro, $20 a month per developer seat. Seen 4 September 2026.
  10. 10Supabase pricing. Pro, $25 a month with the first project included and 100,000 monthly active users. Seen 4 September 2026.
  11. 11Resend pricing. Pro, $20 a month for 50,000 emails. Seen 4 September 2026.
  12. 12Sentry pricing. Team, $26 a month on annual billing. Seen 4 September 2026.
Zac Santer

Zac Santer

Co-founder and design lead, Clientflow

Zac leads design at Clientflow. He scopes every build and writes the number on every quote.

All guides