Nothing is free. The question is what currency the bill gets paid in. For most of the web apps you use without paying, the bill is paid in some mix of attention, data, conversion friction, and time. This is not a moral judgment — some of those bills are perfectly reasonable to settle, and the apps doing the settling have built real value in exchange. It is, however, useful to know which bill you are paying, because the answer changes what you can sensibly trust the app with.
There are, in 2026, roughly four ways a free web app makes money. Most products use a mix. The dominant one tells you almost everything you need to know.
Model one: advertising
The oldest free model. You see ads. The advertiser pays the platform. Display ads, sponsored placements, programmatic auctions feeding a banner you barely register. YouTube, Spotify's free tier, almost every news site, most of the long tail of utility websites — these are advertising businesses.
What it costs you: attention, and the kind of personalization data needed to make the ads more valuable than blanket ones would be. The exact cost has gotten harder to measure since the third-party cookie was finally killed by browsers in late 2024, but advertising platforms have not collectively gotten smaller. They got better at making first-party data work harder.
What it does to the product: it makes the product optimize for time-on-site. Every advertising-funded product is, structurally, an attention business with a side hustle in whatever it claims to do. This is fine for entertainment products. It is a quiet poison for tools you use to get work done, because the incentives slowly bend the design toward 'keep them here longer' instead of 'help them finish faster.' This is why nobody serious uses an ad-supported document editor.
Model two: freemium
You get a free tier with enough functionality to be useful. Heavier users hit a wall and convert. Notion, Linear, Figma, GitHub, Vercel, almost every modern productivity tool. The free tier is a marketing channel, the paid tier is the actual business.
What it costs you: occasional friction at the conversion boundary. If the boundary is well-placed — meaning, the free tier really is useful and the paid features are genuinely for heavier users — this is the healthiest business model on the list. The company makes money. The free users get genuine value. The incentives align: the product wants you to succeed on the free tier, because successful free users are the source of conversion.
What it does to the product: it focuses development on the things that matter to paying users, which are usually larger teams, more storage, more automation, more administration. Solo and small-team features can stagnate. A good freemium company will fight this. A weak one will quietly let it happen, then express surprise when the free tier feels less polished than it used to.
Model three: data
The harder one to talk about, because the companies running this model rarely say so plainly. The product is genuinely free. The revenue comes from selling, brokering, or otherwise commercially exploiting the data the product collects. Sometimes this is identifiable. Sometimes it is 'anonymized,' a word that does less work than people assume.
The clean cases are rare. Most apps that do this also do advertising or freemium and use the data revenue to fund the rest. The classic warning signs in 2026 are still useful: vague privacy policies that reserve every possible right, partnership lists that read like they were written by an attorney trying to cover everything, and a refusal to commit to a specific list of data buyers.
What it costs you: this is the model where the answer is genuinely 'more than you think.' Even if the data is benign, the supply chain of who eventually receives it tends to grow over time, and the original commitments tend to weaken in updated terms-of-service revisions that arrive in your inbox as bullet points you do not read.
Model four: loss-leader
The cleanest of the lot, and the rarest. A larger company offers a free product because doing so strengthens the larger business in some other way. Google Maps is free because Google is an advertising business. Google Docs is free because it pulls people into the Google Workspace orbit. Apple iCloud's free tier exists because it makes the surrounding hardware stickier. Microsoft Teams was free for years because it kept people away from Slack.
What it costs you: dependency on the larger strategic logic. As long as the free product is strategically useful, it gets funded and improved. When the strategic logic changes — and it always does, eventually — the product is downsized, deprecated, or quietly retired. Loss-leader products are the most likely to disappear with a polite blog post in five years.
The cost nobody talks about: time
Every free app has a switching cost paid in time. If you use it for two years and then it shuts down, gets acquired, or pivots away from your use case, you owe a migration. The migration is real work. It is, very often, more expensive than a year of the paid version would have been.
This is the most underrated cost of free-tier dependence. Pick a free product casually, get attached, build workflow around it, and then realize four years later that the migration to a replacement will eat a week of your life. That is also a bill. You will not see it on the invoice, but you will pay it.
What we look for when we recommend free apps
Three rough heuristics:
- 1Can we identify the business model? If we cannot, we are suspicious. Companies that do not know how they will make money tend to figure it out at the worst possible moment, in whichever direction is cheapest for them, not for users.
- 2Is the free tier likely to survive a strategic shift? Freemium tiers usually do. Loss-leader products usually do not. Data-funded products are the worst case — they tend to either disappear or quietly degrade.
- 3Is the data exit story honest? If we can export, migrate elsewhere, and not lose the work, the free tier is mostly fine to lean on. If the export is hostile or absent, the apparent freeness is a long-running bill on a delay.
When 'free' is actually free
Open-source projects funded by community contributions or institutional sponsors are often the genuinely free option. Excalidraw is the obvious example on dotstore — there is a paid tier for teams, but the core product is real software with no advertising, no data harvesting, no conversion wall on the basic workflow. Inkscape, Penpot, Joplin, Cryptpad — there are real open alternatives in most major categories now, and the gap to the commercial leaders has gotten smaller every year.
Open source is not a panacea. The pace of development is slower. The aesthetics are sometimes a generation behind. The support is community-driven. But the cost is honestly free, and that is rare enough on the modern web to be worth noting.
A practical test
Before you sign up for a new free web app, spend forty seconds on three questions:
- How does this company make money? (If you cannot answer in one sentence, that is information.)
- What is the conversion wall? (If there is one, where? If there is not, see above.)
- If this disappears tomorrow, what do I lose? (Decide whether you can afford that.)
Forty seconds is not a lot. It is enough to stay out of most of the bad bets. The good ones still pass the test.
The price field on every dotstore catalog entry tries to be honest about which model the app actually runs on, not just what the marketing page claims. If you spot a case where we got it wrong, write in — we will fix it and credit you in the changelog.
Tags