The Blog

  • December 12, 2013
  • Get a Horse

    many-asses1I was literally gobsmacked and I had to re-read the post several times.  Gartner analyst Robert Desisto—who I don’t know at all—wrote a short post last week saying that today’s SaaS vendors, “will resist the move to ‘pay as you go’ because it will have a very big impact on their business model predictability” and become “legacy dinosaurs” as his headline said.

    But, but, but! I stammered to myself.  How can that be?  I have been researching and writing about this space for fourteen years.  I was the first analyst to cover Salesforce and a bunch of other early entrants, and one of the first people to have a practice dedicated to SaaS.  They all had pay as you go models, at least back then.  Did I miss something?

    One of the real challenges of running a subscription business, and this includes SaaS companies as well as the Dollar Shave Club, ZipCar, and all the other companies that jumped on the bandwagon, is that you have very different revenue flows that must be accounted for.  Companies like Zuora have built big businesses and attracted hundreds of millions of dollars in venture funding to build billing, payment, finance and accounting systems that cater to this massive industry.  Now along comes Gartner with the clear implication that the pay as you go model is not in fact alive and well?  I didn’t get it.  Still don’t.

    As a sanity check I contacted Tien Tzuo, CEO of Zuora, a subscription billing, payments and finance provider.  In his previous life Tzuo was CMO of Salesforce and at one point had the job of inventing a billing system for Salesforce that operated the way subscriptions run.  Here are some points from Tien.

    • Just because some SaaS companies do three-year contracts that doesn’t make them enterprise software dinosaurs.  Every successful SaaS company realizes that keeping churn low is a core part of the model, and every successful SaaS company realizes that long term contracts do not equate to low churn—the only thing that truly reduces churn is to have strong adoption and customer success.  That’s why SaaS vendors invest in customer success while on premise software companies do not
    • Many SaaS companies actually don’t offer three-year contracts.  At Zuora, we see lots of companies with month-to-month models.  CDNs, cloud companies, API companies, point-of-sale systems—these industries all skew towards month-to-month.  Radian6 also had a month-to-month model.  The post also says doing three-year contracts makes SaaS companies vulnerable to other startups who choose to offer month-to-month … but there’s nothing to stop the SaaS vendor from changing their billing policy whenever they want. (my note: provided they have a product like Zuora that makes this easy to do the billing and accounting).
    • Customers don’t have to accept three-year contracts.  It’s naive to say that it’s the SaaS vendor that forces it on them—many companies actually prefer long term contracts once they are committed to the SaaS vendor, as this gets them the best price as well as longer-term price protection.  This can be a win-win scenario.
    • This does create havoc on revenue recognition.  Monthly billing makes billing messy but revenue recognition easier.  Annual or multi-year billing makes billing easy but revenue recognition very hard.  There’s no free lunch.

    It was such an odd thing to read.  It reminds me of some other chestnuts like, “If god wanted man to fly he would have given us wings,” or “We will never need telephones in England because we have such an abundant supply of messenger boys,” or “Someday every town will have a computer,” or my favorite, “640 KB is all the memory your computer will ever need.”  These are all such Luddite comments you just knew upon hearing them that they won’t stand the test of time.  Heck, this one didn’t survive a day before people started scratching their heads.

    Perhaps the last word on this comes from the most authoritative source—the marketplace.  On December 10, BrainSell, a Boston-based technology company announced it would offer an integrated solution of Intuit’s QuickBooks with bi-directional synch to Salesforce.  According to the press release, “What’s really great is that customers can get a Salesforce subscription from BrainSell with no contract, and the ability to pay month to month!”


    Published: 10 years ago


    • December 23rd, 2013 at 9:36 am    

      You do have a point that SaaS vendors like Salesforce have been on a ‘pay as you go’ model relative to what the market was doing before cloud delivery became the norm.
      However, as customers have become more empowered and are pushing the business-to-individual commerce model, they are now demanding a more granular sales approach that goes beyond a yearly, monthly, or even weekly, subscription.
      There are examples like AWS ( or CloudPassage ( who sell down to very incremental levels – truly embracing the concept of selling to an individual person or organization’s needs.
      In doing this, those two companies are making ‘older’ SaaS vendors look a little archaic, but, more importantly, they are avoiding the lead nurturing “purgatory” that prevents customers from buying SaaS products because the ‘minimum’ for a sale is too high for them and thus are never seen as ready for interaction with direct sales.

      • January 2nd, 2014 at 5:36 pm    

        Hi Michael,

        And happy New Year! Thanks for reading and I get your point. Now that I’ve had a little time to reflect on this I have come to a new realization — all SaaS is not the same and one of the reasons we have this discussion, which resembles a lite beer commercial (Tastes Great! Less Filling! I just love this example) is because the two sides are talking past each other. Who do we think we are, Congress? Here’s the thing. For some subscription services the more atomic you can be the better. So delivering an atomic product at a very low cost makes all the sense in the world. As a practical matter, though, there are times when the atomic idea is pure theory and it will never be put to use (see my post on Hypothesis and Theory, you’ll like it). An enterprise buying SaaS services will not likely be sweating the Nth user and it seems to me that my piece and the Gartner post are focussed there. Tease it all apart and you and I can agree that we’re both right without copping out because we fail to find absolute truth. At any rate, I get you and thanks for writing and reading. D

    • December 12th, 2013 at 2:03 pm    


      I’m scratching my head also, so I can only conclude that Rob Desisto meant something other than what we interpreted him to have written. Perhaps he meant that SaaS companies are resisting month-to-month, or even day-by-day adds and deletes. That was’s original model when I got there, but it was so confusing for the customers to get debit and credit invoices every time they added or deleted a license, and it was so onerous on SFDC’s accounting department, that we changed the rules and built the billing system that Tien and others were involved in. Customers had the equivalent of a billing day, and any licenses they added during mid-period were charged at the full price, and they weren’t allowed to reduce licenses except on contract renewal date.

      So perhaps Desisto means “pay as you go” to mean adding and subtracting licenses every day, which approaches the “utility” computing model, but at tremendous cost in friction.


      • December 12th, 2013 at 2:48 pm    


        Could be though in the short piece he didn’t get into any of that. It would be nice to get a clarification from his but for now it’s it and that’s that.

        Happy Holidays!

        PS Don’t forget to vote in CRM Idol!

    • December 12th, 2013 at 11:24 am    

      While I did not read the research note from Robert, do have a couple of thoughts to share. While I agree with the point he was apparently trying to make, I’d also point out that there are a lot dinosaurs flying around this planet. Suspect Gartner may have run into some vendor clients who were perhaps over reacting and stubborn — it’s been known to happen…. May also be another example of trying to just be seen through the noise out there and challenges in the consulting and service cluster. Suggest all analysts take a scroll through AWS to see just a small sample of the variety of models today–only a representative sampling–small percentage, but a good one being used at scale. MM – Kyield

      • December 12th, 2013 at 2:49 pm    

        I like that! The post asks more questions than it answers. D

    Speak Up

    You must be logged in to post a comment.