Don't Offer a Free Plan

Somewhat tautological perhaps, but also, thereby, not a compelling reason to offer a free plan, nor to be proud of being glib.

In addition highlighting that any comparison between an influencer and a consultant depends on false equivalence, which is why I don't consider them part of a product ecosystem (esp. when compared with ISVs, VARs and MSPs), or a substitute for thoughtful, professional advisors.

> Would you rather have 10 users that pay $100 monthly, or 1,000 users that pay $1?

I think this question is meant as an illustration but perhaps it hints at a rule of thumb. Let me substitute a different question and see what you think:

Do you want 1000 users paying $500 a month, or 5 million users paying a $1 a month?

5 million suggests a broad but shallow appeal to the app, and I don't doubt at all that the author has seen customers exhibit all of the listed behaviors - and more. But perhaps an application with a narrow but deep appeal makes sense with a free tier. You can always expand the appeal of your product to reach more markets once you have a proven track record and word of mouth working for you.

I would also say that at a minimum, if you don't let people test drive your application, you have to be quite skilled at creating demos and walkthroughs, and encourage third parties to do the same. Growing that kind of talent in house is probably too slow and unreliable. You're going to have to pay for at least one person to give you good advice.

How did it come to this? Well, at least in the short term, this kind of business busts up traditional high-quality paid tools. If a Google-like business enters a market where a high-quality paid competitor exists, they will subsidize something similar with VC money and suck users away, bankrupting the competitor. Then, with the power they have gained they are free to focus on monetizing their users, which is a story for another time.

This is especially made easier by the fact that it is really difficult to get paid on the Internet: you are essentially selling access to information, not a physical good (all of this might not apply to businesses rooted in the physical world like Amazon and Airbnb). When an average user sees a free service and a paid service both offering what is essentially information, they will go free every day. Since the price of reaching people on the Internet is so low, and the difficulty of getting paid is high, freemium platforms emerge and can outcompete paid competitors by leveraging this fact.

As a consequence to this, quality tools have retreated into paid niches, driven to raise prices. We are essentially left with two product types: gigantic Google-like freemium platforms, and high-quality high-expense alternatives (e.g. Feedly in the Google Reader example). While the former might suck at serving their users and providing value/quality, they make so much money that a cheap, quality tool or a paid, large platform will be quickly snuffed out like a candle without huge investor subsidies.

Well, and saddening that this drivel seems to be a reliable way to get on the HN home page.

1. It’s rare for a business to experience dominantly viral, product-driven growth. Most don’t, including the successful ones! The few businesses that do succeed that way tend to attain ubiquity, so people mistakenly assume a free tier is the best path. (This winner’s bias must be a known fallacy, but if so I don’t know its name.)

2. Moving churn earlier in the customer journey speeds up iteration cycles and increases fidelity of feedback. We’ve all seen “hugely successful” apps that get a lot of users but never find a business model. Maybe if they started charging earlier in the customer journey they would have found product-market fit. It is extremely difficult to calculate customer value and pay for marketing without rapid feedback.

3. Not charging users can create hideous feedback loops. Just take a look at Facebook.

4. Free tiers are a consequence of the non-linear reality of venture-backed teams. Backing capitalists prefer a 5% chance of scoring 1,000 than a 50% chance of scoring 10, so they will of course encourage freemium (or rather, metrics like “daily active users”) alongside a whole host of other too-risky-for-most methods.

5. The author is completely correct that those willing to pay are the easiest to serve, and those unwilling to pay are most likely to need help, leave poor reviews, and give misleading feedback.

6. Again as highlighted by the author, the infrastructure behind multiple tiers is way more costly than you’d expect. It’s easy to get caught in the business mechanics when one could be making the product and marketing better. How costly do you think it will be? Multiple that by 3 to get the actual engineer/design hours, now multiple it by 3 again to understand how distracting it can be.