I should really about this in a more formal fashion, but here's the HN comment version of several years of doing pricing for SaaS companies:
Price higher than you feel comfortable with. Approximately every SaaS team devalues their software because, unlike their clients, they were actually capable of writing it. Your clients don't care that it was "only a few weeks of work" or "really not as good as $UNRELATED_SOFTWARE_YOU_COMPARE_IT_TO" or "not as polished as Apple." Also, if you haven't run a business before, you have no idea how much businesses pay for pedestrian services like e.g. trash takeout, business insurance, monthly bookkeeping, etc.
Concretely, my standard "knowing nothing else" pricing for B2B SaaS is a three plan grid with $49/$99/$249 . If you're servicing informal firms like e.g. many Shopify sellers, you can add in a $29 price point. You don't want to service businesses for below $29 a month; you will suffer enormous pain in doing so and you will find that they churn and burn through software for a variety of reasons unrelated to you, for example because they go out of business, they have such severe cash-flow constraints such that sending you an email asking for a $5 discount is a good use of their time, etc.
You're not going to get pricing right the first time around. No one does. This is fine. Over time, you are likely going to raise prices across the board, as you improve your product, get a tighter focus on which customers benefit from you (which is isometric to improving the product, from the customer's perspective -- c.f. a below comment from tptacek on packaging), and get more confident in your team/offering. There is a simple way in SaaS to make price increases a non-event: grandfather in existing customers at their existing price. If your company is growing, your revenue for the next month is dominated by existing clients but your revenue as T goes to infinity is dominated by new clients, so I tend to just recommend people grandfather indefinitely. Yes, this does result in a couple of early adopters getting $1k a month of services for (in some cases) $19; I consider that a marketing expense to reward people for loyalty.
There are a lot of microtactical things you should build to support experimentation with prices. One is backend infrastructure to allow your CS team to change someone's plan at any time; another is some sort of crediting feature or special one-off pricing (one-time or ongoing), because as T goes to infinity you'll probably want to do both of them. I'll write about them some day.
You can track churn rates by the plan people are on, but I find that that isn't terribly useful early in the life of a business, since you'll have relatively few accounts in any bucket. For this reason, I often have a shadow attribute on plans to do "bucket by relative size of account", such that e.g. folks on Small Business at $49, Small Business (March vintage) at $34, Starter Edition For YC Companies at $45, etc all bucket into the same place for churn calculations.
Prefer giving people free things (exceptional acts of service, for example) to giving people discounts. They'll remember them for longer and you absorb the costs once, early in the relationship, rather than every month. Any recurring discount you issue in SaaS is essentially a liability on your books. (And existing happy accounts are a de facto asset, though GAAP doesn't treat them that way.)
Thanks for the comment. But Github (to take one example) has pricing that starts from $7 a month. This despite of having a free tier. What's your opinion on that? Is it because they target developers and not businesses at that price point? Even their organisation plan starts at I think $11.
If Github priced it so low, I guess they wanted to capture that segment of developers who would pay $7 a month but no more. It makes me want to price my product too at $7 a month at a lower tier. Any comments on this please?
I think GitHub intends to have approximately every software developer in the world use GitHub, which is not the planned endgame for most software companies.
I do not think, if your customers number less than hundreds of thousands, that getting the custom of folks for whom $9 is a lot of money is a worthwhile use of your time. Devs consistently overestimate the impact on their marketing/etc of having cheapskates use them, and underestimate the real operational costs. (These customers are disproportionately pathological; they contact support rudely, with inane questions, at a rate orders of magnitude higher than good customers.)
The other thing is the $7 price point is for an individual / enterprise. You can't have a team collaborating which means that isn't the price point any business is paying.
Price higher than you feel comfortable with. Approximately every SaaS team devalues their software because, unlike their clients, they were actually capable of writing it. Your clients don't care that it was "only a few weeks of work" or "really not as good as $UNRELATED_SOFTWARE_YOU_COMPARE_IT_TO" or "not as polished as Apple." Also, if you haven't run a business before, you have no idea how much businesses pay for pedestrian services like e.g. trash takeout, business insurance, monthly bookkeeping, etc.
Concretely, my standard "knowing nothing else" pricing for B2B SaaS is a three plan grid with $49/$99/$249 . If you're servicing informal firms like e.g. many Shopify sellers, you can add in a $29 price point. You don't want to service businesses for below $29 a month; you will suffer enormous pain in doing so and you will find that they churn and burn through software for a variety of reasons unrelated to you, for example because they go out of business, they have such severe cash-flow constraints such that sending you an email asking for a $5 discount is a good use of their time, etc.
You're not going to get pricing right the first time around. No one does. This is fine. Over time, you are likely going to raise prices across the board, as you improve your product, get a tighter focus on which customers benefit from you (which is isometric to improving the product, from the customer's perspective -- c.f. a below comment from tptacek on packaging), and get more confident in your team/offering. There is a simple way in SaaS to make price increases a non-event: grandfather in existing customers at their existing price. If your company is growing, your revenue for the next month is dominated by existing clients but your revenue as T goes to infinity is dominated by new clients, so I tend to just recommend people grandfather indefinitely. Yes, this does result in a couple of early adopters getting $1k a month of services for (in some cases) $19; I consider that a marketing expense to reward people for loyalty.
There are a lot of microtactical things you should build to support experimentation with prices. One is backend infrastructure to allow your CS team to change someone's plan at any time; another is some sort of crediting feature or special one-off pricing (one-time or ongoing), because as T goes to infinity you'll probably want to do both of them. I'll write about them some day.
You can track churn rates by the plan people are on, but I find that that isn't terribly useful early in the life of a business, since you'll have relatively few accounts in any bucket. For this reason, I often have a shadow attribute on plans to do "bucket by relative size of account", such that e.g. folks on Small Business at $49, Small Business (March vintage) at $34, Starter Edition For YC Companies at $45, etc all bucket into the same place for churn calculations.
Prefer giving people free things (exceptional acts of service, for example) to giving people discounts. They'll remember them for longer and you absorb the costs once, early in the relationship, rather than every month. Any recurring discount you issue in SaaS is essentially a liability on your books. (And existing happy accounts are a de facto asset, though GAAP doesn't treat them that way.)