Hacker News Reader

Comment by tlb | Hacker News Reader

While the title sentence can be true under steady-state conditions, it's not true for virally growing startups. Growth depends on customers recommending the product to their friends. For example, if raising prices by 10% cuts your monthly viral spread coefficient by 10%, it will cut your market share by 90% after 22 months.

(I'm working on a simulation engine for this kind of stuff.)

View on HN

The opposite is true for normal startups. Unless you drop the price to zero, the viral coefficient goes up with price. The people willing to pay the most are the most likely to spread the word. You will have fewer customers at a higher price, but the likelihood any given customer will spread the word is higher because you are solving a more important problem for them and they are more committed to the product.

If your SaaS startup meaningfully depends on viral growth, and a plan that's 10% cheaper has an effect on it, you'll almost always do far better with a freemium business model.

It's hard to recommend freemium for most SaaS startups since it depends critically on having a huge potential market and a zero-cost avenue for growth within it.

Yes, but market share isn't everything. If you're making 10% more per customer with 10% less customers, you're worrying about keeping less customers happy and can focus on making the smaller group even happier for ~ the same revenue. I'd be interested in seeing your simulation if you decide to make it available.

"(I'm working on a simulation engine for this kind of stuff.)"

How do you calibrate it for a specific market?

There's no silver bullet. Version 1 will at least let you capture your assumptions about how you think your market works and how your product will get traction in it. As you gain experience in the market, you'll be able to refine the model and make better comparisons among alternative growth strategies.

Eventually I'll try to curate a library of market models and startup plans that did or didn't work in them.

But what drivers do you model? Consumer behavior and preferences, or just abstract business metrics? I.e.: do you put into your model what is typically an indicator on a dashboard - TLV, retention rate, etc, or do you derive a model of your customers from past data and build scenarios of changes in customer behavior?

Reply on news.ycombinator.com