The five core pricing models for a subscription business are a regular fixed-fee subscription, advertising, e-commerce, transaction processing, and data and APIs. Most strong businesses mix two or more, with a subscription as the backbone. Since 2022, usage-based, credit, outcome-based and hybrid pricing have joined them, especially for AI products.
Key takeaways
- The goal of every pricing model is to move customers toward recurring revenue.
- Advertising and transaction processing only work at scale. Build them into another model rather than relying on them alone.
- Freemium needs an enormous addressable market. If your free version strips out what makes the product worth having, use a free trial.
- Monthly billing lowers the barrier; annual billing lowers churn. Offer both, with an annual discount you can afford.
- In 2026 the most common shape is a hybrid: a subscription base with usage, credits or outcomes on top.
Michael Martin stood at the bottom of an escalator in New York's Grand Central Station, handing out business cards to everyone who came up.
"I was like, 'Yeah! We're going to have a million users after today!'" he said later. "So, I finally take the escalator up myself after I'd given them all out, and three-quarters of them are in the trash bin at the top of the escalator."
Martin was a co-founder of RapidSOS, an emergency technology startup. The cards promoted Haven, an app that let you push a button and send your location and critical details straight to a 911 call handler. Simple, clever and potentially life-saving. It had won about $500,000 in startup competition prize money and $60,000 on Kickstarter.
The plan: a six-month free trial, then $2.99 a month.
People didn't want it. Downloading an app for an emergency felt like too much effort compared with dialing 911. And why pay $2.99 a month for something they got free when they called?
Haven was shut down in 2018. RapidSOS then changed its model completely. Instead of charging consumers, it built technology that lets companies connect their apps and devices to the 911 system. When I wrote the book, it was processing more than 150 million emergencies a year, had 350 million connected devices and had raised hundreds of millions of dollars.
Same mission. Same team. Different pricing model.
That's why this matters. Even a product with a life-saving benefit can be rejected if the pricing model doesn't connect with the people paying. And if your pricing doesn't land, the moat you've built won't help you, because nobody will be inside your castle to protect.
What are the main types of pricing models?
You already know the model I recommend. It's in the name. Subscription.
But subscription isn't the only option. (If you haven't picked which of the eight subscription models you'll run, start there.) You can mix it with other models to build a stronger offer. In the book I settled on the five most useful:
| # | Pricing model | Who pays | Works best when | Main risk |
|---|---|---|---|---|
| 1 | Regular fixed-fee subscription | The customer, on a schedule | You deliver continuous, visible value | Customers can't see why they'd pay every month |
| 2 | Advertising | Advertisers, for your audience's attention | You have a large, engaged audience | Small audiences earn almost nothing |
| 3 | E-commerce | The customer, per purchase | You sell products people want to own | One-off sales, no recurring base |
| 4 | Transaction processing | A small slice of each transaction | You facilitate high volumes of payments or deals | Pennies per transaction; needs huge scale |
| 5 | Data and APIs | Companies that use your data or plug into your platform | You hold data others need | Privacy rules and proving usefulness |
Since the book was published, four newer models have become mainstream, mostly because of AI. I cover them in the 2026 update below and on their own pages: usage-based, credits, outcome-based and hybrid.
Model #1 – Regular fixed-fee subscription
I know what you're thinking.
"Rob, you've already waxed lyrical about subscriptions. Do you need to go over it again?"
Yes. Mostly because I want to show you subscriptions work in places you might not expect.
Take cars. Zipcar, founded in 2000, charged a small monthly fee (about $7 when I wrote the book) for on-demand access to a car, plus an hourly rate per drive. It's a fixed subscription with a pay-as-you-go layer, perfect for people who want a car without the cost of owning one.
Or airlines. United Club membership gave access to lounges worldwide and other perks for $650 a year when I wrote the book. (Not a bad deal if you can put up with their customer service. Sorry, United.)
The advantages run both ways. You get predictable revenue that doesn't depend on one-off sales. Customers get a set fee they can budget for.
The caveat is the RapidSOS lesson. A subscription only works if customers can see the benefit of paying regularly. If you're not delivering continuous, tangible value, they'll leave in droves.
What has changed since 2022
The book also quoted Jaguar's car subscriptions at about $1,000 a month, including registration, insurance and maintenance. That market has largely retreated. Volvo suspended Care by Volvo in the US on 1 August 2024, and Jaguar has paused most new car sales while it relaunches as an all-electric brand. When a fixed fee has to cover a depreciating asset, insurance and early returns, the maths gets hard.
Meanwhile, fixed-fee subscriptions have kept their pricing power where customers depend on them. Spotify raised its US Individual plan from $11.99 to $12.99 in January 2026 and still grew to 306 million Premium subscribers by the third quarter of 2026.
Model #2 – Advertising
This model is simple. Instead of customers paying, advertisers pay for the attention your service attracts.
Mobile games do it well. PickCrafter, developed by Fiveamp, used rewarded video ads: instead of interrupting play, it offered in-game rewards to players who chose to watch an ad. Revenue grew 165 percent and ad engagement rose 94 percent. Players got to enjoy the game without constant interruptions.
The martial arts gym that sold ad space
You don't need to be a media company. One of my clients was a martial arts gym in the US. I helped them partner with local businesses (a Domino's, a FedEx and a bakery among them) to create a "Neighborhood Discount Card." Each business offered discounts to the others' customers and got featured on a flyer that sat on every participating front desk. Everyone got more exposure for free.
If you run an SEO agency, you could sell placements on your site to web designers and PR firms your clients also need. You build partnerships and earn revenue from an asset you were only using for marketing.
The big caveat
Advertising only works when enough people show up. A YouTube video watched by a thousand people probably won't make a dollar. Podcasts nobody listens to don't attract sponsors.
It's all in the numbers. You can also mix it with subscription by charging a monthly fee for a specific ad placement.
What has changed since 2022
Advertising has made a comeback inside subscriptions. Netflix's cheaper ad-supported plan reached more than 250 million monthly active viewers by its May 2026 upfront, although Netflix cut its 2026 ad-revenue forecast from about $3 billion to about $2.7 billion. OpenAI now sells an $8-a-month ChatGPT Go plan with ads. The pattern: an ad tier gives price-sensitive customers somewhere to go instead of canceling.
Model #3 – E-commerce
E-commerce accounted for more than $4 trillion of sales in 2020 alone, the figure I used in the book.
Catch of the Day, and what happened next
Catch began in 2006 as Australia's first "deal of the day" retailer. Some of its early days were remarkable: 2,000 Asus laptops sold in an hour, and $1.5 million of Samsung TVs between 12 p.m. and 1 p.m. It grew into a marketplace with more than two million products and sold to Wesfarmers in 2019 for $230 million.
It also had a paid membership, Club Catch: $69 a year for free shipping and exclusive deals, with more than 250,000 members at the time of sale. Members spent more per basket and bought more often. Wesfarmers rebranded it as OnePass and added free delivery from Kmart and Target.
Here's the update. In January 2025 Wesfarmers announced that Catch would stop trading as a standalone business in the fourth quarter of its 2025 financial year, moving its fulfillment centers to Kmart Group. OnePass carried on, described by Wesfarmers as "driving incremental sales" across its retail brands.
Read that again. The e-commerce business closed. The membership survived. That's the book's argument in one sentence.
Amazon: the ultimate moat
You can't talk about e-commerce without Amazon. It sells its own products, takes a fee from third-party sellers, sells ads on its own pages, runs Prime and runs Amazon Web Services. It combines every model on this page. Amazon's "subscription services" revenue, mostly Prime, was $13.73 billion in the second quarter of 2026 alone.
The weakness of plain e-commerce is my old agency problem: one-off sales. Without another model, there's no guarantee a customer returns. That's why you build in subscription. LARQ sells a self-cleaning water pitcher upfront, then ships replacement filters every two months on subscription. That's e-commerce and consumables working together.
Model #4 – Transaction processing
Escrow.com does exactly what its name says. It holds money and goods until both sides of a deal are ready. One seller of the domain marijuana.com used it to let the buyer pay over forty years, with Escrow.com collecting the payments along the way.
Transaction processing is about being the facilitator and taking a small slice each time.
PayPal is the obvious example. In 2020 it earned about $21.4 billion of revenue on about $936 billion of payments, a little over 2 percent, and $4.2 billion of profit. Only (and I use that word loosely) $4.2 billion on almost a trillion dollars of volume.
You need huge numbers and deep pockets. Your first years are about investing and scaling until volume turns a profit. That's why I see it as a model you build into another one. Amazon and eBay charge sellers a fee on every sale.
What has changed since 2022
Stripe shows how powerful the model is at scale. Businesses on Stripe processed about $1.9 trillion in 2025, up 34 percent and about 1.6 percent of global GDP, and the company was valued at $159 billion in a February 2026 tender offer.
Small marketplaces have learned to use transaction fees more cleverly. StyleSeat, the salon booking platform, now charges stylists a monthly subscription plus a one-time "New Client Connection" fee of 30 percent of a new client's first appointment, capped at $50, and never again for that client. Charging only for the first transaction removes the incentive to take repeat clients off the platform.
Model #5 – Data and APIs
Remember RapidSOS? It failed with a consumer subscription and succeeded with an API that other tech companies plug into. That API moves data to emergency responders.
Data businesses sell data, usually as a service. Customers choose what they need, often pay only for what they use, and sometimes pay a subscription for access. (See how the models mix again?) CoreLogic sells historical and current property data for an annual fee to banks, insurers, government and individual investors.
Two challenges. Data is only valuable if it helps the buyer reach a goal, and companies are already drowning in their own. And privacy rules are tightening. People get touchy about businesses selling their data.
Frankly, so do I.
What has changed since 2022
AI has made data and APIs one of the fastest-growing models. Google's Looker now meters its conversational analytics in tokens, at $3 per million input tokens and $20 per million output tokens, with monthly allowances by edition. Arena, one of 2026's fast growers, turned a free AI leaderboard built on more than 10 million user votes into a paid, consumption-priced evaluation service for AI labs. Developer infrastructure companies such as Exa and Deepgram publish simple per-unit API rates with free starter credit.
You don't have to stick to one model
I've said Amazon is my idea of the ultimate subscription business, because it combines all five models. Plenty of other great companies combine several.
Spotify is the classic. Free listeners hear ads, so Spotify earns from advertisers. Paying subscribers get no ads and more features, and that subscription is the backbone of the business.
For years Spotify was "great product, no profit." That changed. 2024 was its first full year of operating profit, about €1.4 billion, and by the third quarter of 2026 it had 306 million Premium subscribers and 788 million monthly users. About 39 percent of its users now pay. The free, ad-funded tier is the funnel. The Spotify case study shows how it got there.
I'll repeat it: you don't have to stick to one model. Use one to feed another.
Find your pricing model in five minutes
Flat, tiered, seat, usage, credits, outcome or hybrid? The free Pricing Model Picker recommends one for your business.
Should you go freemium or offer a free trial?
On the surface, freemium and free trials sound alike. They're very different.
- Freemium gives a limited version away forever. You push free users toward paid, and you may earn from free users another way (Spotify's ads).
- A free trial gives the full product for a limited time. Then customers pay or stop. Streaming services have used trials for years.
Both can be gateways to a subscription. Which you choose depends on two things.
1. The balance between value and comfort
Freemium needs a free version good enough to be worth using, and a paid version with enough extra value to convert a decent share of users. Spotify's free tier still lets you listen to plenty of music; Premium removes ads and adds offline listening and other perks.
If limiting your product ruins it, use a trial. Think of Netflix offering a free tier that doesn't let you watch movies. Nobody wants that.
2. Your addressable market
Jason Lemkin, co-founder of EchoSign, shared some maths that every founder should see.
Worked example: the freemium addressable market conundrum
Say you charge $10 a month and you want $100 million-plus of annual revenue.
You need about 1 million paying subscribers: $10 × 1,000,000 × 12 = $120 million a year.
Freemium conversion rates are commonly quoted at 2 to 5 percent. At a conservative 2 percent, you need 50 million free users to get there.
Your market has to be big enough for that before freemium even makes sense. If it isn't, raise the price (and deliver the value to justify it) or use a free trial.
Canva shows freemium working at that scale: more than 265 million monthly users and more than 31 million paying users in 2025, about one in nine. Most businesses don't have a market that size.
| Choose FREEMIUM if… | Choose a FREE TRIAL if… |
|---|---|
| Your addressable market is enormous | Your addressable market is small |
| Your product is naturally shareable | Your product isn't naturally shareable |
| A free version is genuinely useful and paid still offers plenty more | A free version would strip out what makes the product desirable |
There's a third option now: the reverse trial, where users get the full product for a limited time and then drop to a free plan instead of being locked out. The full comparison is in freemium vs free trial.
Trials and the 2027 Australian rules
From 1 July 2027, Australia's new unfair trading practices law is expected to require businesses to remind customers before a free or discounted trial ends. As at October 2026, the detail is still being finalized, so treat this as general information and check your billing system can send trial-end reminders automatically.
Annual vs monthly subscriptions: which should you offer?
Each has its pros and cons.
Annual billing guarantees a twelve-month relationship, gives you time to prove long-term value and makes churn easier to manage. It's also simpler to collect one payment than twelve. But it asks customers for more money upfront, so some won't sign up.
And every twelve months there's a big renewal event. One of the best salespeople I ever worked with called it "poking the bear." If the customer is happy, why make them decide again every year?
Monthly billing removes the commitment barrier, which is why about 70 percent of subscription services bill monthly. The price looks smaller, so value is easier to show. But customers can cancel at short notice, churn can bite before you recover acquisition costs, and admin costs rise.
Many companies offer both, with a discount on annual. Panera's drink subscription is a neat 2026 example: $14.99 a month, or $119.99 a year. Twelve monthly payments come to $179.88, so the annual plan is about a third cheaper, which pulls members into a 12-month commitment and brings cash forward.
How big should your annual discount be? It depends on your monthly churn. Our guide to annual vs monthly plans shows how to work it out, and the calculator below does the maths.
Annual Discount CalculatorFind the annual discount that still beats monthly churn.Open the free tool →2026 update: the new pricing models
When I wrote the book, I called usage-based pricing "the emerging pricing model" and gave it its own chapter. Four years on, it has emerged. Here are the models you need to add to the original five.
| New model | How it charges | 2026 example | Best for |
|---|---|---|---|
| Usage-based | Per unit consumed (API call, message, gigabyte) | Twilio per message; Snowflake credits for compute | Products whose cost and value scale with use |
| Credits | A prepaid allowance spent on actions of different sizes | GitHub Copilot's monthly AI credits (from 1 June 2026); ElevenLabs credit tiers | AI products where tasks vary in cost |
| Outcome-based | Per result delivered | Fin (formerly Intercom) at $0.99 per resolution | Work with a clear, countable result |
| Hybrid | A base subscription plus usage, credits or outcomes | Anthropic's $20 Enterprise seat plus usage | Most software in 2026 |
The data backs this up. ICONIQ's 2026 survey of about 305 AI builders, as summarized by SaaStr, found consumption pricing rose from 35 to 42 percent of companies and outcome pricing from 18 to 23 percent, with the average company blending 1.7 models.
Each has its own page: usage-based pricing, credit-based pricing, outcome-based pricing and hybrid pricing.
Whichever you choose, the hardest decision is the unit you charge on. That's your value metric, and it should grow as your customer gets more value. Our guide to choosing a value metric walks through it with examples.
How to choose your pricing model
I can't tell you the perfect combination for your business. That's up to you. But I can tell you two things about pricing.
First, whatever models you use, move customers toward subscriptions. Recurring revenue creates predictability and makes your business more attractive.
Second, keep it low friction. A confusing pricing model means fewer people pay. A smart, simple pricing strategy can speed up your growth and widen your moat. Sometimes the quality of your pricing, not just the price, becomes part of your competitive advantage.
And be honest about when a model is wrong for you:
- Don't rely on advertising if your audience is in the thousands, not the millions.
- Don't build a transaction business unless you can fund years of scaling.
- Don't go freemium in a small niche. Use a trial.
- Don't meter usage if your customers' usage barely varies. A flat fee is simpler.
- Don't sell data you haven't got clear permission to sell.
What to do next
- Write down the pricing model that feels best for your business. Mix and match if you need to.
- Run the Pricing Model Picker to test your instinct.
- Decide freemium, trial or neither, using the table above.
- Set your monthly and annual prices, and check the annual discount against your churn.
- Read the rest of our subscription pricing strategy hub for tiers, packaging and price rises.
Read the whole book free
The Subscription Playbook as a free ebook and audiobook. Australian business owners can claim a free hardcover, shipped free.
Claim the free hardcover Get the ebookFrequently asked questions
What is the difference between a pricing model and a pricing strategy?
A pricing model is how you charge: a monthly fee, per transaction, per use, through advertising and so on. A pricing strategy is how you set the level and structure of those prices: value-based, good-better-best tiers, discounts, price rises. Choose the model first, then use strategy to set the numbers and packaging.
Which pricing model is best for a small business?
For most small businesses, a regular fixed-fee subscription with two or three tiers is the safest start. It's easy to explain, easy to bill and gives predictable cash flow. Add usage charges only if some customers cost far more to serve than others, and add advertising or transaction fees only once you have real scale.
Can I change my pricing model later?
Yes, and many successful companies have. GitHub moved Copilot to usage-based credits in 2026 and Clay split its pricing into two units the same year. Change openly, give notice, show customers a worked example of their new bill and consider keeping existing customers on their old terms for a period.
What is a reverse trial?
A reverse trial gives new users the full paid product for a limited time, then drops them to a free plan rather than locking them out. It combines the 'feel the full value' benefit of a free trial with the long-term funnel of freemium. It suits products with a large market and a genuinely useful free tier.
What is a good free-to-paid conversion rate?
The book quotes the commonly cited freemium range of 2 to 5 percent. Results vary widely by product. Canva reported about 31 million paying users out of more than 265 million monthly users in 2025, roughly one in nine. Free trials that ask for a card usually convert at much higher rates than open freemium plans.
Should prices include GST in Australia?
When selling to Australian consumers, the price you advertise generally needs to include GST and any compulsory fees. Business-to-business prices are often quoted excluding GST, but be clear about it. This is general information as at October 2026, so check the ATO and ACCC guidance or ask a tax adviser for your situation.
Get both books free
The Subscription Playbook (ebook + audiobook) and Feed a Starving Crowd, instantly. In Australia? Claim the free hardcover.
Robert Coorey, MBA
Robert is the author of The Subscription Playbook and Feed a Starving Crowd and a Director at BPAG Accountants in Sydney. More about Robert.
Sources
- Spotify: Q3 2026 earnings (22 Oct 2026)
- Variety: Netflix ad tier reaches 250 million monthly active viewers (May 2026)
- Wesfarmers: Catch wind down and OneDigital update (2025)
- Stripe: 2025 annual letter (24 Feb 2026)
- Google Cloud: Looker pricing (accessed Oct 2026)
- Intercom (Fin): pricing (accessed Oct 2026)
- TechCrunch: Canva gets to $4B in revenue (18 Feb 2026)
- SaaStr: 10 metrics from ICONIQ's 2026 State of AI report (10 Jul 2026)
Facts and prices checked 27 October 2026. Spotted something out of date? Tell us.