Thursday, June 21, 2012

The Price is Right (or not)

One of the most important questions which every SaaS company has to solve is to find the right pricing – the right pricing model as well as the right price levels. It’s obvious that getting pricing right is extremely important: If you’re too cheap you will leave money on the table and reduce your ability to invest in customer acquisition. You may also hinder adoption especially from bigger customers who think that your product can’t be good because it’s so cheap. If you’re too expensive you might be scaring away the majority of your potential customers.

It is of course impossible to find the optimal price point in a way academic textbooks would define it. Finding that would require you to do more tests than you can possibly do. What you should do is try to get to that point as close as possible, and when I talk about the “right” pricing I mean a reasonably right pricing.

Unless your target customers are all very similar (which is unlikely), the most important thing that your pricing model has to accomplish is to capture different amounts of money from different customers based on their willingness and ability to pay, which correlates with the value that they’re getting from your product. In the old enterprise software world this used to be the job of the sales people – talk to the customer, find out about his needs, get a sense for what he can pay, offer him a solution and negotiate a price. In the world of SaaS, customers (rightly) expect more transparency and will look for a price list on your website before they start a trial.

In many cases a per-user pricing (often also referred to as “per seat”) is an obvious choice, and some of the most successful SaaS companies including Salesforce.com are using that. Other successful examples include pricing based on:
  • number of clients managed with the software (e.g. Freshbooks)
  • number of newsletter emails sent (e.g. MailChimp)
  • number of email recipients in the system (e.g. ConstantContact)
  • amount of storage that is used (e.g. Dropbox)
  • number of events tracked (e.g. KISSmetrics)
What these companies have in common is that they've found an "axis" that highly correlates with their customers' willingness to pay, which allows them to keep their service affordable for small customers while asking bigger customers for much more. It also allows them to benefit from the growth of their customers, since a growing company needs more seats/emails/MBs/events/etc over time. Ideally this can lead to what is known as "negative churn" – the wonderful situation when the MRR growth of some customers of a customer cohort more than offset the effect of terminations from that cohort.

Importantly, most successful SaaS companies differentiate their prices along more than one axis (David Skok wrote about this here). Secondary axes include the level of support, additional features or other usage parameters. For example, Freshbook's pricing is based on a combination of the number of clients that you can manage and the number of seats, plus two additional factors:


So what's the right pricing model for your SaaS startup? For obvious reasons it depends and I have no general answer to that question, but here are a few practical tips:

  • Try to find one or more axes which correspond with the value that your customers are getting from your product and which correlate with your customers' willingness to pay. Talk to your customers and analyze how your early users are using the system to find out the ways in which larger customers are using your product differently from smaller customers.
  • In the beginning, err on the side of being too cheap rather than being too expensive. In the beginning the most important thing is to get customers. You can optimize your margins later.
  • Later on, make sure you're not leaving too much money on the table. If not a single customer ever complains that you're too expensive that's a strong sign that you're too cheap.
  • Accept the fact that it's very unlikely that you will get your pricing right at the first shot. Go out with something that you think makes sense, get feedback from the market and be prepared to make changes quickly.
  • If you increase prices, try to do it along with new value-add features that help justify the price increase. And offer your existing customers extremely generous grandfathering terms.
  • If your pricing is differentiated based on features, consider giving all users the high-end plan with all features during their trial so that they can play around with the full product.
  • Maybe not necessary to mention since these are all known best practices, but just in case: Give users a self-service free trial. Offer monthly pay-as-you-go subscriptions that users can cancel at any time. Provide an option to pay in advance for a year (with a discount). Create a clean, beautiful pricing page. 
Do you know any examples of pricing models that worked or didn't work, or would you like to get my feedback on your pricing model? Get in touch!

8 comments:

  1. Hi Christoph,

    nice article, but for finding the right price it is maybe a bit to skindeep...their is a reason, why big strategy consulting companies like SKP are existing...but basically a good beginning at all...pricing i s also what i am most interested in, so if you may like to do some research in this field...i would be happy to work with you...

    ReplyDelete
  2. Hi Markus,

    Thank you for your comment. You're right, I have of course only scratched on the surface of this topic. Just some tips and tricks for startups – who can't afford hiring a strategy consulting firm for their pricing strategy! :)

    Best regards

    Christoph

    ReplyDelete
  3. Anonymous3:46 PM

    Every of your 'practical tips' (thank you) has included the most important thing not only for pricing, but too for running the business successfully: The importance to focus on real customer feedback. Not only for deriving an improved pricing structure from out of your observances.
    I suggest one more way of dissecting the problem of the 'right pricing', somehow making a prior step.

    For the very first clue on how much to ask for, I think it's usefull to reconsider two different things separately: Every customer, no matter weather he was a private consumer or a b2b customer, makes a buying decision with a more or less structured underlying decision procedure. The decision is determined by his needs/ sentiments/ thoughts/ goals..., but for your 'very first clue', I think that all of your customers aspects come together in two criteria:
    (1) By how much money is your product helping your customer saving costs, and
    (2) How much are you able to implant a 'must have hype' into your customers mind.
    Now let's think about these two aspects separately:

    (1) The more precise you are able to calculate the monetary advantages of your deals offered, the more precise you might guess how much to ask for. A good relation might be fifty-fifty. If each deal managed by your system is saving your b2b cutomer 20 mins of time for 2 coworkers (causing costs 120.00 an hour each), then ask your customer for 40.00 for each buying contract (of the over all 80.00 of his total manpower cost). If on top of this, all your b2b customers by now are paying their tax counsellor 1.00 for each data row (contract) while preparing their tax returns, they might be willing to pay you 0.50 if the usage of your system prevents them from hiring a counsellor for preparing their papers for the tax office. If on top of this...
    You guess what! Think beeing your future customer and calculate the monetary advantages of using your system. And your marketing can argue, that it's only your 50% of the bargain, the rest is his!

    (2) Many businesses sell a feeling, a hype, and even some feed themselves by abusing their customers imagination beeing 'a part of the pack'. They do not sell a 'monetary use' (think of cars, swimming pools, lingerie, ordinary gad-gets ... social networks, ring tones for your mobile ... allmost everything). The worth for your customer by beeing marked as a 'part of the hipsters' is hardly calculatable. Nevertheless, there might be a lot of comparable offers out there on whiches pricing structures you might leer at.

    There are more aspects on the 'right' pricing (e.g. think 'turnover vs. market share', ...), but finally your pricing structure will become a mix of (1) and (2). And thinking in the terms described, will open you another way in finding the right ratio for each consumer cluster.
    Our western societies are still on the move towards a more and more shimmering world in which even the construction worker no more accepts to drive an oily muddy bulldozer. Glam is increasingly important and aspect (2) still increases in its participation.

    C.U., Nik

    ReplyDelete
  4. Anonymous3:36 PM

    Totango.com published a report some time ago where we researched 550 SaaS companies and found that 41% of these companies publish pricing on their website. Of those 46% have per user pricing, almost all of which scale in more than one dimension (David Skok or Salesforce.com style). The median pricing for an entry level feature package was $25 per user and the highest value feature package was $275 per user. Of course the spread was quite large from $1 per user to $15,000 per user depending on the application. You can read more here: http://blog.totango.com/2012/02/freemium-free-trial-and-pricing-models-in-550-saas-companies/.
    I very much agree with your two main points:
    - Price according to the value you deliver to your customer/user
    - Start low, deliver value first, and grow the revenue per customer/user over time
    That does however require of course that you put in place business processes by which you systematically move users to successful implementations and to increase their usage of your product over time. You can read tons about how to do this on our blog of course blog.totango.com.

    ReplyDelete
  5. Anonymous3:37 PM

    Totango.com published a report some time ago where we researched 550 SaaS companies and found that 41% of these companies publish pricing on their website. Of those 46% have per user pricing, almost all of which scale in more than one dimension (David Skok or Salesforce.com style). The median pricing for an entry level feature package was $25 per user and the highest value feature package was $275 per user. Of course the spread was quite large from $1 per user to $15,000 per user depending on the application. You can read more here: http://blog.totango.com/2012/02/freemium-free-trial-and-pricing-models-in-550-saas-companies/.
    I very much agree with your two main points:
    - Price according to the value you deliver to your customer/user
    - Start low, deliver value first, and grow the revenue per customer/user over time
    That does however require of course that you put in place business processes by which you systematically move users to successful implementations and to increase their usage of your product over time. You can read tons about how to do this on our blog of course blog.totango.com.

    ReplyDelete
  6. Christoph,
    In your opinion, is per user pricing always the best default way to price? Or are there axis which are more appropriate for the particular business?

    Also, with a per user price, what is stopping users from gaming the system? eg. multiple users using the same account? What have you found?

    ReplyDelete
  7. Jindou, thank you for your comment!

    I think the best pricing model always depends on the particular product/market but I really like per-user pricing since it scales very well from smaller to bigger customers.

    I don't think that cheating is a huge problem. Usually it's just too cumbersome to share logins.

    ReplyDelete
  8. How do we create a pricing model for some thing like Amazon AWS?

    For example, my startup generates code for Business Apps that work on Web,Mobile etc. Each app is different.(Made by users them self) And I am struggling to find a simple pricing model.

    Also I need to host that app. So if so bad pricing will affect me every month.

    Any suggestions?

    ReplyDelete