Freemium Mechanics

I read an interesting blog post by Ruben Gamez titled Why Free Plans Don’t Work. If you are interested in freemium business models or any of the variations on the theme, this is well worth reading however I take issue with a couple of points.

First and foremost, Gamez uses a statistics breakdown (in %) to highlight the disparity between free and paid plans. Whenever someone does this they invariably open the door to the question about what their customer numbers because a percentage breakdown without knowing what the denominator is will lack the proper context. Knowing that 1% out of 100 customers are paid versus 1% out of 100,000 is a fundamentally different discussion to have… and there is no discussion about the cost to serve free product customers.

Gamez points out a number of well known freemium companies and the transitions that they have made between free products and free trials. This is an interesting discussion and the body of work that can be studied is relatively small and fluid given the immaturity of freemium as a business model. However, a couple of things are increasingly apparent for people who are running these businesses.

You can have a freemium business that depends on a free trial process instead of a free trial and a free product option at signup, there is no debate about this, and you can have an exclusively paid product that depends on a free trial process for acquisition and onboarding. This is a smart decision in my opinion and at Get Satisfaction we are constantly tinkering with and evaluating the options relative to placement and purchase path for the free product. The idea here is to route every website visitor who becomes a prospect into a funnel that exposes them to the full product before downgrading them to a free product.

In 2010 we relaunched our website with a new “plan picker” page, which over the course of the year went through 2 significant updates that are very relavent to this analysis. Initially we had Free placed as a promo box on the sidebar, separate from the monthly subscription plans but highly visible nonetheless… this is the control group as best I can provide one because with each subsequent change to the plan picker page we changed more than just Free product placement.

In April of this year we elevated placement of the Free product to equal standing with the monthly subscription plans. Almost immediately the number of new communities created through the free product jumped substantially (and for the record, I am not going to disclose actual customer numbers so I’ll do my best to avoid putting up percentages, following my own advice above). At the same time the number of new trials created for our monthly subscription products remained flat and in some months declined materially, however the number of free-to-paid conversions for customers who were net new (not a previously paying customer who canceled) went up.

The net result was still a decline in new customer conversions and our churn rate (turnover of all paying customers in a single billing period) stayed constant or declined slightly so I would have to say that elevating Free to first world status did not improve the business.

In August we changed the plan page again and pretty much hid the Free product option. The resulting decline in Free product signups was dramatic but offset by the trial signups and associated trial conversion rate, however churn went up as well so the net effect was offset by customer cancelations in the first 90 of total life.

Churn is a really important consideration in freemium models and not just because of the financial impact. The raw churn number is obviously important because that represents the size of the hole you need to fill each month before you can start adding customers, however when churn happens is often overlooked.

You should be doing a cohort analysis each month on cancelations to determine what the survival curve is for each customer segment, which graphically represents how quickly cancelations are happening in the customer lifecycle as represented by the 25th, 50th, and 75th percentile groups.

This first example is basically a bad curve because it shows that over a proscribed period of time a large percentage of your customers fall off. It’s basically telling you that you are attracting the wrong kind of customers and you are going to invest disproportionately in replacing lost customers.






This next curve is a pretty good one, the drop is initially steep but then levels out and after 12 months you still have over half of the customers you acquired in any single cohort. What this curve is telling you is that you are losing customers who are not a good fit for you very quickly and then cancelations stabilize.






In the context of freemium this information is very valuable because it is a consequence of how prominently you are positioning free vs paid product options. If you are hiding free in order to stimulate take-up rates on paid, then you have to expect that cancelations rates will go up as a result of people converting to paid that otherwise would not if presented with a prominent free option.

This leads to the next topic I want to discuss, which is the methodology you embrace for the trial process. In the interest of being honest and transparent, the way we do it at Get Satisfaction is not the optimal way to do trials because we provision trials as a time based variant of a specific product instead of having a single trial where everything is turned on and then have the prospect select the product they want to convert into at the end of the trial process.

The second problem we created for ourselves is that we require the web visitor to create an account and give us their credit card information in order to create a trial account. This is an obstacle for trial creation first and foremost but also orients the trial experience to people who are pre-disposed to buying you before they even enter the trial process… so in effect you are giving them a free period of service for something they would pay for.

We are going to make changes to the trial process to address the two issues I raise, however I can’t do much about the account creation requirement simply because my product requires a named user to be the administrator of it… no user registration would mean I would have no account to attach the administrator rights to. My recommendation to you is that you create a free trial process that downgrades to a free product at the end of the trial period if someone doesn’t enter their credit card details and select a plan, instead of offering a trial experience in addition to a free product.

Annual billing options are a game changer in the freemium model, arguably the single most effective strategy for reducing your churn rate. Typically the way that annual billing is presented is 12 months of service for the price of 10, a 16% discount.

You can also use a buy-it-now option to bypass the trial process, offering something like a discount or promotional offering in order to pull forward demand that exists in the trial pipeline, and in the process isolating true prospects who are won or lost in the trial. I’d like to do this at Get Satisfaction as part of our structural changes to the trial process.

Having a freemium business model is dependent on a number of strategies but one that often gets overlooked is how well you identify potential demand and feature marketing inside the free product for free to paid conversion and inside the various monthly subscription products for paid-to-higher-paid conversion.

Ultimately the freemium model is a strategy that increases the catchment of leads as a result of using your product as the primary marketing vehicle through which you deliver a funnel to. Take care to structure your website so that every aspect of the content you are creating is designed to deliver a site visitor into a product experience or isolate them for followup through a traditional enterprise sales process.

It’s also worth pointing out that if you have a product that you are primarily selling to businesses, and the product itself has a multistep onboarding process, then you really have to have a higher touch sales process where you are nurturing the free and trial accounts at a higher level than if you were, for example, Evernote.

For me the mechanics of a freemium business are some of the most interesting to be involved with in a modern software as a service company. The implications of billing and provisioning system dynamics, how you structure your website content, surface funnel analytics, build upselling cues into your application, and manage high volume sales nurturing processes are incredibly complex but increasingly normal for the B2C and even B2B markets.

CLV, Cohort Analysis, and Survival Curves

Customer Lifetime Value (CLV) is an important metric in any SaaS business, it explains the potential for profitability and more directly the extent to which a company should spend to acquire customers, measured as Customer Acquisition Cost (CAC).

Custora is a company that provides an analytics platform specific to customer value and survival, which is a cohort analysis of customers acquired month by month to analyze how quickly each group churns out.

CLV forms the foundation for a probability model through which a company can forecast future recurring revenues, because every SaaS company exists within the frame of every customer renewing their contract at the end of each billing period (month).

The cohort analysis weights the group of customer acquired in a given month with their probability for renewing. Think of it this way, every customer has a coin they flip at the end of the month, renew or cancel, and in any given population of customers you want to know what the probability weighting of coins are for renewal; the cohort analysis provides this weighting while the CLV estimates the value of an individual customer in that population.

The customer survival curves are valuable for a couple of reasons, the first being the rather obvious benefit of seeing where most of your churn is happening, however this leads to a more insightful analysis about the type of customers you are attracting as a company. Simply put, if the customer survival curve is very steep, as in a fast falloff as churn chips away at each cohort, then the conclusion you can draw is that you are attracting the wrong customer for your product.

Churn is a naturally occurring factor in all SaaS businesses (and all non-SaaS businesses!). The point is that you have to strategize around managing it down to a number that is equal to or better than the average for your peer group and analytics such as what Custora provides are a good investment to make in this area.

At Get Satisfaction we have used a variety of homegrown and data services to instrument our business for day-to-day management and long term planning. We worked with Custora under their previous company name and then I let the service lapse as I built up an internal data warehouse of performance tracking metrics. I missed the dashboard and analytics drill downs that Custora provides, which because they are very focused around a specific problem set tend to be more useful than endless spreadsheets, so we are back to using Custora for customer retention analytics.


SAP’s John Schwarz on Analytics

This is Part 2 in my Fortune Brainstorm Tech series. Part 1 covered a conversation with Vishal Sikka.

Vinnie Mirchandani, Oliver Marks, Dennis Howlett and I sat down with John Schwarz, who recently came to SAP via the Business Objects acquisition and has quickly emerged as the man with the plan. If you have been following SAP’s financial results this year you already know that Business Objects is fueling SAP’s growth as their core ERP business has been hit by the global economic recession and stalled.

That analytics is fueling new business growth is not surprising, SAP’s business has historically been driven by transaction and system of record applications (think of all the large three letter acronym categories) but as those systems mature the demand for upgrades subsides and as a consequence of having been at this for so many years there is simple marketplace saturation and “going downmarket” hasn’t exactly worked out very well because SaaS applications have emerged as preferred solutions for small and mid tier customers.

Fast forward to today and analytics is looking to be a really attractive market for SAP because of the strength of the Business Objects offerings and good integration with SAP’s application products. Competitors include IBM and Oracle but neither of those vendors are particularly well positioned because IBM is selling database oriented solutions and Oracle hasn’t achieved broad integration of the analytics assets they have acquired, so they are selling pieces and parts rather than a solution.

Schwarz is responsible for this group within SAP and I think it’s pretty safe to say that they have integrated the company and their 6k employees, or at least have done the bulk of the integration heavy lifting. The question that now exists is “where are you going to take this thing now that you have the keys and a full tank of gas?”.

The answer is two pronged, the first being that Schwarz makes a tacit acknowledgement that most of what all these companies have been selling in years past is more “reporting” and less “analytics”, and even in that frame customers have not demonstrated great competence at utilization of these capabilities. Another reality is that the best systems have been applied against cubed data (OLAP) systems as opposed to real time operational systems, and by definition has been targeted at specialized use cases (like how your credit card company analyzes potential fraud).

Schwarz’s answer to this is moving the state of the art on two separate axis, the first being “analytics for everyone” by delivering solutions that have a very refined user experience for reporting, ad hoc query, and historical and predictive analytics. The second axis is potentially more disruptive to the market but one that will most certainly present more headwind in the form of mature competition that SAP will have to overcome while on a steep learning curve, and that is unstructured data (free form text).

The first axis, analytics for everyone, has high probability for success because SAP & Business Objects are really well positioned to expand their footprint in customer sites by going direct to the business user, who according to Schwarz is the decision maker for analytics purchases as opposed to IT. Schwarz and Sikka both referred to the “consumerization” trend in enterprise apps and it was refreshing to hear them say it because this has been a central theme in my writing for several years now and it is simply undeniable, business users expect business applications to be like the applications and services they use in the consumer world.

The move into unstructured data is more problematic for SAP on three fronts. First and foremost, SAP doesn’t have any packaged data feed bundles they can sell and relying on partners has proven problematic for other companies in this space. I asked Schwarz if this suggests that SAP will be acquiring wholesale content providers that they can layer on top of applications and he acknowledged that it is currently a “raging debate within the company”.

While I think SAP should acquire a content aggregator and perhaps even someone like Gnip, I doubt they will do it, instead going to companies like Factiva and Thomson Reuters to provide bundled content for SAP customers and in the long run they will come to the conclusion that trying to be a neutral player while bringing together content providers is simply cumbersome and foists upon the customer too many contingencies. However, I will acknowledge that given the state of their progress on this initiative, acquiring a content provider now does not give them a strategic advantage therefore time is on their side.

A second concern I have is that unstructured text analysis is very difficult to do well and even specialist companies that have been at this for years find it difficult, so to suggest that SAP will simply do this because they set their mind to it is optimistic to say the least. Entity extraction, semantic tagging and triple building, sentiment analysis, quality, and content authority are extremely challenging, especially as content feeds shift to activity streams which provide far less context for the content to be analyzed against. To then marry this to structured data analysis is something that has been accomplished in highly specialized fields like government intelligence and applications for financial traders. Don’t get me wrong, all of the things I have just written about are exciting and well worth doing, I simply think you have to be prepared to fail more than you succeed, even if you are as accomplished as Business Objects is.

Lastly, there is one very big obstacle in the way of anyone attempting to capture unstructured data for analytics, office applications. The single greatest store of unstructured data in any enterprise is email, followed closely by documents created by personal productivity applications. Microsoft is certainly well positioned here, their entire Sharepoint roadmap reads like a plan to integrate MS Office data to collaboration and transaction systems but even Microsoft is finding that bringing intelligence to email is a tough rock to push. Similarly, companies like Gist, Xobni, Clear Context, Kwaga, and Postbox have emerged and will certainly develop leadership long before SAP enters the market.

So what can we conclude from all of the above, well I think it’s pretty clear that Business Objects under Schwarz’s leadership is well positioned to execute on the vision he has established, from the standpoint of having a solution, market credibility, and access to a substantial customer base, but it’s also clear that they will have to make some acquisitions to fill gaps and build some durable partnerships with companies that matter.

On the partnership front, we spent quite a bit of time talking about the widely known issues that SAP has with the idea of partnering, and by extension the Palo Alto versus Walldorf mentality that has marked their expansion into Silicon Valley. Schwarz insists that these issues have been resolved but I remain unconvinced for many reasons, but mostly because NIH is such an integral part of SAP’s culture that I don’t think they will ever shed it. Secondly, as accomplished as John Wookey is, I don’t believe he can be the force at SAP that he was at Oracle because to do so would involve a massive realignment of turf that would strip many in Walldorf of their fiefdoms and we all know that at it’s core power in any organization is a function of turf.

Personally I think that SAP should stop having an inferiority complex about partnering and instead just focus on building the best tools for partners to take advantage of. Every company that sells into the enterprise wants to align with SAP at some level (hell even Oracle does to support their database business) so instead of trying to sell everyone on “hey look at us, we’re cool like Google and we want to be your friend”, just go out and build kick ass tools that partners can use to build great solutions and then get the hell out of the way.

On the competitive front Schwarz had some interesting things to say about known competitors IBM and Oracle, but his most insightful comments were saved for Microsoft and Google. On Microsoft, Schwarz confirmed that they are looking at Azure with great caution but generalized Microsoft’s competitive weakness as “trying unsuccessfully to be all things to all people”. It’s hard to disagree with that statement, nor to ignore the fact that Microsoft has not been as successful with applications as they would have predicted so for SAP the truth is the devil you know is better than the one you don’t.

Regarding the devil they don’t know, it’s clear that Google has emerged as a competitor for SAP to be concerned about. Schwarz’s general view seems to be that Google doesn’t have the maturity and product management expertise to be successful in the enterprise, pointing to companies like FAST and Autonomy doing enterprise search far better than Google, despite the fact that search is the crown jewel of the Google empire. From where I sit, the single biggest threat that Google presents is their ability to climb the learning curve very quickly and that the lack of rigid product management has demonstrated to be a remarkably effective quality for quickly iterating products that then gain broad consumer acceptance.

Whether or not that agile process is repeatable in the enterprise is debatable but it’s clear that for all the talk about innovations at SAP, the fact remains that we see very little of it in the product pipeline. This isn’t due, IMO, to the stuff not working but rather the tendency for SAP development groups to set the bar impossibly high for a product to be delivered. While I don’t think Google will present a serious threat to SAP for many years, I do believe that SAP can learn a lot from Google that will make them a better company as a result.

To conclude, I really enjoyed meeting Schwarz and it’s clear that he has a good line of sight on where his business unit needs to be focusing not just for revenue growth but also to protect the flanks as the market continues to evolve. While there is much to like, there is also much to be concerned about but on balance the concerns are more tactical than strategic so I’m willing to give him the benefit of doubt at this point.