Why you should not hard code non-linear pricing

Take a look at this very nicely done price estimation tool of Dropbox business edition.

Dropbox Single User

The entry price is $795/year – either you want 1 or 5 users you pay that single price. Clearly it makes no economic sense for a 1,2 or even 3 member team to pay for this version. However if there exists such a team that values the features available only in the Business edition they would pay. But Dropbox decided not to unbundle it any further and make $795 the entry price for a 5 member team. Makes very good sense as it assures a floor price and makes sure the revenue is aligned with fixed cost to serve these business customers.

Then it gets tricky and complex. Having adopted  $795 as the base price Dropbox has created multiple price tiers based on different number of users. While it possible to license for any arbitrary number of users past 5, the tiers act as price boundaries. After 25, 50, 100 etc. users the price per user drops.  Take a look.



The price per user that for all practical purposes starts at $159/user starts sloping down and reaches the lowest value of $127/user with three other price points in between ($132, $128 and $127).

This is classic example of non-linear pricing where the total price does not increase linearly with units but curves down as volume increases. You have seen this from cereal packages you buy to the group discounts you get when you register for conferences. Non-linear pricing is great but it introduces operational complexity to marketer and cognitive complexity to customer.

For what these costs mean let me take you back to my article from four years ago – 4 Costs of Versioning. The many different choices you offer to customers at different price points introduce four different costs (some on you and some on your customers). These versioning costs should not be confused with costs to make and serve the customer. These costs are incurred in completing the transaction.

If the non-linear pricing requires two (or three) price tiers then most of the versioning costs are manageable. But  after two tiers the  costs to customer starts to go up. As a refresher customer cost means-

This is the cost incurred by your customers in understanding all your many different versions to make a choice. These are also the costs you have least control over. The costs may not be incurred in the form of dollars but there are definitely cognitive costs and opportunity cost to the customer. Worse, the effects of these costs do not end after version selection.

I do not mean here Dropbox should not practice non-linear pricing by reducing unit prices for larger customers. There is no need however to programatically (buying program that is) hard code the multiple price tiers to explicitly state the non-linear price curve. Far from helping customers with that decision it will only complicate their choice.

Dropbix could easily provide better pricing to these high volume customers if the customers were to call and ask for it(or provide the sales discount hierarchy for the sales team). That is leave it up to the customers to ask there by making it a better form of price discrimination. Why give programmed discount when customers do not ask? Why complicate pricing with all the hardcoded discounts?

As a contrast I would like to point you to another business that is very close to Dropbox – Box.net (for business customers both Dropbox and Box.net are trying to address the same jobs the customers are trying to get done and hence they are  alternatives).


Box pricing starts at $45 base ($15/user but the minimum users is 3). This is very similar to Dropbox’s minimum limit of 5 users.

Before you compare the two prices please note Dropbox prices are listed on yearly basis while Box lists per month price. Regarding which option is better, that is for another day.


After the minimum, Box provides very simple scalable model of $15/user/month up to 500 users. No other pricing tiers in between keeping it simple. But I bet if you are signing up 200 users you will negotiate a better price than $15/user. It is just not hard coded non-linear pricing.

If we did per month pricing for Dropbox, the price difference from non-linear pricing does not even seem significant for all its versioning costs.

When it comes to non-linear pricing, by all means do it as long as it keeps your multiple editions simple to manage and simple for customers to choose. Anything more than 2, don’t hard code. Leave it up to your customers to ask for it.


If you are selling Enterprise Apps you don’t start with freemium

I thought the word freemium went the way of singing fish and MySpace and hoped I never have to write yet another article with this portmanteau in the title. Unfortunately wrong ideas  and false beliefs don’t die easily. They are not replaced by some profound truth because the believers suddenly achieve self-realization. As Kathryn Schulz wrote in her book, Being Wrong, bad ideas die hard because they can only be replaced by another equally bad idea. Until another such bad idea comes around we are stuck with freemium.

This time we are presented with some profound advice on go to market strategy for Enterprise Apps by Scott Irwin from Rembrandt Venture Partners. In his article for GigaOm , Mr. Irwin recommends freemium as the first option for go to market strategy for Enterprise apps before inside sales and before enterprise sales.

Those who are already sold on the idea of freemium will see this as further evidence supporting their case.  Those who are new to the idea will likely see the popularity of the post as evidence for its veracity. Those like me are not going to be convinced as usual. The problem this time is the flagrant errors in the case Mr. Irwin makes by recommending freemium for enterprise apps.

If you stopped reading here, think about it – Enterprises have a budget and have wherewithal to pay. Why shouldn’t you charge for your value-add?

Now to the flaws in Mr. Irwin’s argument.

  1. Ignoring Customer Needs: There is absolutely no mention of the customer segmentation and their needs. Why are customers hiring the Enterprise 2.0 Apps for? If you do not understand your target segment and their needs you cannot deliver them an effective product. And if there is an urgent your product fulfills why should you not charge for it? These are enterprise customers and they have a budget to pay for these apps that add value.
  2. Ignoring Customer-Channel Alignment: Mr.Irwin starts out by making a case with Salesforce.com, a company I admire for its disruption of the enterprise software landscape and its marketing. But it should be noted that they very carefully chose their initial go to market strategy that aligned with how enterprises buy software – building an highly effective enterprise sales team backed by phenomenal marketing. It was not freemium that helped Salesforce.com grow to $3 billion a year company. Sure their product was easy to setup and use but they were not just fighting against customer apathy, they were competing against strong players with significant sales prowess. Do not for a second think freemium would help compete against entrenched players or serve as free marketing.
  3. Choosing Irrelevant Examples: If the topic is about go to marketing for Enterprise Apps the examples used should at the very least use such companies. Not Evernote, a consumer based webapp. When it comes to freemium examples, for the past two years, there have been no other examples than Dropbox and Evernote. Such a model of try the free version and upgrade to premium may work in consumer segment (barely, only 3% upgrade to paid version) but the competition is not going to let that happen for enterprise segment.  In addition any such popular example also suffers from biases.
  4. Choosing Selective Evidence:  Mr. Irwin makes a case using SurveyMonkey, specifically goading us to make app fun so users will use it. First, why should making the app fun be mutually exclusive to charging for it? What about many other applications that are fun to use and but not free. If we want to stick with the same application family as SurveyMonkey, we have SurveyGizmo which you know decided against freemium model to target enterprise customers. There are many other examples of applications that are fun to use and not free. By using selective evidence Mr. Irwin not only succumbs to biases but leads his readers down the wrong path.
  5. Anything but Charging For Value: Rest of his article is presented as a recipe for freemium. If you did not have your segmentation right, you do not have your product strategy or go to marketing strategy right. Any other revenue model, however innovative it is, is not effective. Yes Atlassian and others adopted pay to charity, pay what you want, pay with WOM etc models. Likely these models were relevant for them because they started with right customer segmentation. But all those do not apply to your business.

Why are management gurus, entrepreneurs, startup gurus and now venture capitalists  dead set against getting fair share of the value they create for their customers?

Pricing Multiple Editions – SurveyGizmo Takes a New Approach

My favorite survey platform is SurveyGizmo. In the past I have written about its pricing and how it effectively used multiple versions and visual nudges in its pricing page. SurveyGizmo has been experimenting with their editions and pricing page since then. From presenting five options, to four options and now there are only three options when you visit their pricing page.

Before I point out the most critical change in their pricing, let us look at some of the secondary changes

  1. What is missing in the three options? What is one version you see in any pricing page you visit but is missing here? The free version. It is not prominently featured in SurveGizmo page. It is still there but as a footnote. It is an indication that their customer mix has changed as they move into next phase of the product adoption.
    Their current customer mix is more likely made of Enterprise customers with willingness to pay for a survey platform and a budget to match it. The focus has likely shifted from attracting freeloaders who may never convert to those who think differently about the product and have different buying process.
  2. What do you see about the prices? The highest priced option is listed first and the middle option is prominently featured (in the middle too). This points more to the size of organizations or groups within organizations they are targeting. While you may notice the two options as different you will later see this difference essentially going away.
  3. What do you think about unlimited number of responses in all three? Most webapps differentiate based on number of responses or equivalent – like number of Giga Bytes of storage in case of Dropbox or number of events per months in case of Kissmetrics). SurveyGizmo has done away with number of surveys or number responses as pricing meter. It is a very good approach as most likely customers are not seeing as many responses and it does not make sense as a meter to attach pricing to.

Now all these points are for naught when you try to upgrade your free account to a paid account. Despite what the pricing page says they have done away with any feature differences between the different editions. In essence there is just one version of the product with all the features.

Well the free edition comes with limitations, otherwise you would be happy with free.  Beyond that are no difference in the power of the tool, types of questions, reports, number of emails you can send, etc.

If they have done away with differences what is the pricing meter then? They rely on number of users. Want access to all these features? You can get it for $50 and after that it is $20 each additional user on the account.

Why have they done away with multiple editions? If one price is good, aren’t two better?

When you have  multiple versions (editions) these should differ in at least two dimensions. The mandatory dimension is price and you choose the second based on what the customer values and willing to pay the price difference.

For example, take MacBook Air. Its multiple versions differ in three choice dimensions. Price, screen size and capacity. Clearly the customers see value difference between 11″ and 13″ screens and are willing to pay for it.

But if the customers do not see value difference between versions, they serve no purpose. In fact they add to cognitive cost to customers in making their purchasing decision. When SurveyGizmo had Personal, Professional and Enterprise editions they tried to limit the advanced features like custom scripts to the certain versions. It is likely that only a small percentage cared about these and for the rest the most essential features of the survey platform were more than enough.

Hence their decision to get rid of multiple versions/plans/editions and charge only based on number of users.

How do you decide on offering multiple versions of your product?

Related Articles:

  1. Why there is only one version of Apple TV but three versions of Roku?
  2. Why are raspberry and strawberry yogurts priced the same?
  3. Should your Versioning differ in quantity or benefits?

Note: I have used words Plans,Editions and Versions interchangeably in this article.