Neil Gupta

The Definitive Guide to Business Models

1232 words • 5 minutes to read

Founders often struggle to pick the right business model for their product. Your business model is the fundamental engine that defines your business and goals. After people issues, I would bet choosing the wrong model for your strategy is the second largest cause of product failures. There are only a few possible options, so let’s discuss1 them in detail from worst to best if your goal is to build a sustainable business.


Let’s start with the ever popular free business model… Giving a product away for free is really the absence of a business plan. You are trading sustainability for rapid growth, and either kicking the monetization problem down the road or monetizing your users in some other way. This is like trying to start a campfire with rocket fuel. Sure, you’ll accomplish the stated goal but probably not get the results you were hoping.

Free is often necessary for social networks that need to overcome the network effects hurdle and can’t extract value from solo users. Unfortunately these networks very rarely succeed at sustainably monetizing later, so their only options are acquisition or failure (raising lots of VC money at high valuations is not success).

Interview vs Job

You are not Facebook.

If you’re debating whether a free model makes sense for you, re-frame the question as whether your business would still make sense if you paid people to use your product, because that is what you’re doing. Many founders delude themselves with complex plans of monetizing after they hit certain milestones, but this is like betting on multiple miracles at once. While giving away products for free is a rapid (albeit expensive) path to growth, remember that the time it takes your startup to reach peak popularity is your half life. Sometimes a slow burn is actually what you need.

Sell attention

This is when you monetize your users’ attention via ads. It’s usually the natural follow-up to free models. There is nothing wrong with selling ads per se, but it requires your app having lots of engaged users first. If you’re picking the model for a new product, you are probably falling for the multiple miracles fallacy (lots of people will use your app, and then lots of advertisers will be interested in your user base). Plus, nobody likes ads, so your incentives are no longer aligned with users either, and these products always fall victim to enshittification eventually.

Here is how platforms die: first, they are good to their users; then they abuse their users to make things better for their business customers; finally, they abuse those business customers to claw back all the value for themselves. Then, they die.

Avoid this model!

Sell privacy

The other way to monetize free products without ads is to directly sell user data to someone else. For example, Edison Mail does this by building a free email app and selling market insights to their actual customers. While this approach is even more likely to rely on multiple miracles, it can be a workable model for certain types of problems.

Sell consumables

Consumables are ongoing purchases that get spent and then need to be purchased again for continued use. This could be razor blades like Harry’s, coffee pods like Keurig, or printer ink like HP. Enterprise support plans and extended warranties also fall in this category. These models can be successful if the consumable itself provides value to users, or abusive if the consumable is forced upon users to justify the company’s business model. This model is also the most susceptible to competitive pressure since users can switch whenever the consumable runs out.

Facilitate transactions

If you own a marketplace, you can facilitate transactions between users where you take a cut of the value created. Stripe, Airbnb, Uber and others do this to great effect. This feels free to users because they are only paying when they make money, and perfectly aligns your interests with users’ interests. This is actually the most preferable model if it works for your business, but it is usually limited to marketplace startups.


When you charge users a subscription to access your product, you are renting the product to them. It’s the most sustainable model for the developer, but also most frustrating for users. However, certain products require a subscription model and should lean into that. Users are ok subscribing if they understand why. For example, Flighty provides high quality flight data that costs them money, thereby incurring an ongoing cost the more I use Flighty.

Sell directly

This is the simplest and oldest business model. You charge users upfront for value and live and die by the quality of your product. The vast majority of products are bought upfront and you can keep using it for as long as it lasts. This model works well for software too, if you charge for major version updates. Charge your users for v1 as you work on the next big thing. When you release v2 next year, it’s totally fair to expect users to upgrade if they want the new features. Those who don’t can keep using their purchased version. Everyone is happy. This works best if your product does not have ongoing costs, like servers or expensive API integrations.

For example, I’ve paid for multiple major Reeder versions (currently on v5) when new features are added, but also know I can keep using Reeder 5 for as long as I want since it has no server component and I don’t expect significant updates to this version. Unfortunately, Bear chose the wrong model for their business. They are charging a subscription by locking Apple’s built-in features like iCloud syncing behind a paywall. There is no ongoing cost for Bear developers, other than future development. While I pay for Bear anyway, this does leave a bad taste in my mouth. If I ever stop paying, I will lose access to features I’ve already paid for for no reason. For their product, it would be far better if they charge me upfront for Bear 2 and when they are ready to release Bear 3, charge me again!

As long as you’re not trying to rent-seek, selling your product should bring roughly equivalent revenue to renting it. If you don’t think you’ll be able to release major updates fast enough to match subscription revenue and don’t have any other ongoing cost, then you probably shouldn’t be charging a subscription either because users are going to expect rapid updates for their subscription.

In other words, if you can, choose the direct sale model. Unfortunately founders are often reluctant because it feels too simple. Well, I’m here to remind you that telemetry only gets you so far. The best way to know if you’ve made something people want is when they hand over cash for it.

  1. We’re talking about products only. Human services, like consulting, are not included. Also, pricing is a tangential topic to model. First you have to pick the right model, and then you choose the right price with that model for your specific product. Pricing is out of scope for this post, but you probably should charge more

Written on August 28, 2023 in Chicago.