Highlights
- Organizations leave an average of 36% of their SaaS licenses unused (Zylo, 2026), and 80% of features in the average software product are rarely or never used (Pendo, 2019).
- Median SaaS spend has reached $9,455 per employee across organizations Zylo tracks. Small businesses run leaner, but the direction is the same: up.
- The reason your software feels too big is that it is: the tools you rent are built for the average of thousands of businesses, and your business isn't average.
- A 15-minute stack audit (below) shows you which tools earn their bill and which ones you're paying to fight.
- The replacement path is real now: 35% of teams have swapped at least one SaaS tool for a custom build (Retool, 2026).
Open your CRM and count the buttons you've never clicked. Somewhere in that toolbar is a forecasting module priced into your subscription and a territory system you'll never assign. You're paying for all of it, every month, per seat.
That's the small business software problem in one screen. The tools aren't bad. They're built for the average of every business that might buy them, and the average of thousands of businesses describes none of them, including yours.
How much of the software you pay for goes unused?
Most of it, by both of the ways it gets measured. At the license level, companies leave an average of 36% of their SaaS licenses unused (Zylo, 2026): seats bought, billed, and never logged into. At the feature level it's starker: 80% of features in the average software product are rarely or never used, per an analysis of 615 products (Pendo, 2019).
The bill for all of it keeps growing. Median SaaS spend has reached $9,455 per employee across the companies Zylo tracks (Zylo, 2026). Those are enterprise-weighted numbers, and a five-person business won't spend like a 5,000-person one. But the mechanics are identical at every size: per-seat pricing, and renewals that arrive whether the tool earned them or not.
Why does rented software fight your business?
Because it was designed to almost fit thousands of businesses instead of fitting one really well. A SaaS vendor building a CRM has to serve the solar installer, the recruiting firm, and the wedding photographer with the same fields and the same pipeline stages. The result is software where your workflow is a settings page buried under everyone else's.
The fight shows up in familiar places: required fields you fill with junk because they don't apply to you, a pipeline that doesn't match how your customers buy, and the spreadsheet you keep next to the tool because the tool can't hold the thing you need most. When you maintain a workaround for your software, you're paying twice: once for the subscription, once in the hours the workaround eats.
Per-seat pricing compounds it. You don't pay for the five features you use; you pay for all 100, multiplied by every person on your team. The price rises when the vendor ships more features you didn't ask for. You're funding a roadmap written for the average customer. You are not the average customer.
What are small businesses doing instead?
Replacing the worst offenders with software that fits. Among teams surveyed in late 2025, 35% had already swapped at least one SaaS tool for a custom build, and 78% expect to build more in 2026 (Retool, 2026). On Bolt.new, custom CRMs jumped from 19th to 8th among build categories year over year, and internal operations tools grew from 8% to 14.3% of everything built (Bolt.new platform data, 2026).
The unlock is that building stopped requiring code. Owners describe the workflow (the fields they use, the report they check on Fridays) and an AI app builder like Bolt.new writes the software around it. The result has exactly the features they use, because those are the only features they described. Seven of them walked through what that looks like in 7 small businesses that stopped renting software and started building it.
How do you audit your own stack in 15 minutes?
Pull up your bank statement and make four columns:
- List every software subscription you find. Include the ones billed per year; divide by 12.
- For each tool, write the features someone on your team touched in the last month. From memory, without opening the tool. If you can't name three, that's the answer.
- Multiply per-seat tools by the seats you pay for, then circle the seats that belong to people who log in less than once a week.
- Mark every tool that has a workaround living next to it: the spreadsheet, the group text, the sticky-note system. Those are the tools that fight you.
The circled seats are cancellations. The marked tools are candidates for replacement, starting with the one that hurts the most. And the number at the bottom of the column is what "the cost of doing business" has been costing your business.
The average was the compromise
Renting software made sense when the alternative was a six-figure development contract. You accepted the 95 extra features and the workaround spreadsheet because building your own version was out of reach. That trade is over, and you don't have to accept the average anymore.
We're publishing a build-vs-buy calculator in September so you can run your own numbers. Until then, the audit above tells you where to start.
