The Great Strava Debate: To Free or Not to Free

I’ve mentioned the idea of ‘running’ office hours before (i.e., running with students while discussing topics related and unrelated to class), so while at this year's MSOM conference, a group of us gathered for morning runs, which was a great way to step out of the routine of talks and coffee breaks.

I realize this isn’t a new idea, and I’m sure many people do it, but it was great to get in some exercise with colleagues while discussing teaching, research, and life in academia. If you’re a former or current student, I should probably say that the main topic revolved around student bashing. If you’re just a fan of this newsletter, you have nothing to worry about.

As we ran, we realized that most of us use Strava to track our runs. Daniel Freund suggested that I write about Strava and their freemium model. His point was that the only reason he uses the paid subscription version was to make sure Strava remains in existence.

I confessed that I use the paid subscription too, although I’m no longer sure why, or in other words, which specific feature was the one that convinced me to pay. I thought it was the option to see running maps in other cities —one thing I enjoy when traveling is going for a run as soon as possible to maintain my running schedule. During my preparation for the Philly marathon last year, I ran in Philly, NY, SF, Stockholm, Copenhagen, Waterford (Ireland), Montreal, Phoenix, and Mumbai. In fact, I’m writing these lines from Guangzhou after a run (the run was recorded on Strava, as the protocol requires).

I also admitted that the app that I like using most to track my running is Ranalyze, and I still use the free version even though I am willing and want to pay for it for exactly the reason Daniel pays for Strava. Unfortunately I haven’t had the opportunity (or the trigger) to do so since the free version is already pretty great!

When asked why I don’t write about this, I shared that I’m not sure I have something unique or interesting to say.

But just like everything else I write about, I start with an idea and then delve deeper beyond trivial, common sense.

So, if I haven’t already bored you, join me down this rabbit hole (you see what I did there: bore vs. bore, like the boring company).

Strava’s Strategy: Balancing Free and Paid Features

I try to refrain from writing about pricing, but in the world of digital services, one of the most critical decisions a company faces is determining what to offer for free and what to reserve for paying customers. This balance, often encapsulated in the freemium business model, can significantly influence a company’s growth and profitability. And I think Strava provides an excellent example of this strategy in action.

For example, Strava offers a blend of free and premium features to its users. The free version includes features such as activity recording, device support, and social networking functionalities. In contrast, the premium version, offers advanced metrics, heart rate and power analysis, training dashboards, route planning, and more:

United States subscribers pay $11.99/month or $79.99/year plus applicable taxes.

In general, this strategy is designed to attract a large user base through free features and then convert a portion of those users into paying customers by offering valuable premium features.

But how does Strava determine the optimal mix of free and paid features? What are the main tradeoffs? Why offer a free version? Wharton doesn’t have a free version (well, we sort of do with our Coursera courses, but I don’t think conversion is the main strategic factor behind it).

So let’s delve deeper.

Balancing Profitability and Growth in Freemium Models

But why offer a free version?

For multiple reasons:

The freemium model may leverage the concept of network effects, where the value of a service increases as more people use it. This is particularly relevant for social networks and platforms, where user interactions and data sharing enhance the overall experience.

Another reason could be related to high switching costs, where users have already loaded data and have gotten accustomed to a product, and realize the value of the product over time.

It could also be that a customer doesn’t really know the value of the product, and needs time to evaluate it, before deciding to pay.

Finally, it could be all of the above simultaneously.

Existing and ongoing research on freemium models provides insights into how companies can balance profitability and growth, especially when their products have both network effects and intrinsic value. Strava, with its dual emphasis on social networking and fitness tracking, exemplifies this balance.

The paper “Designing Freemium: Strategic Balancing Growth and Monetization” by Lee, Kumar, and Gupta, highlights the balance in freemium models between offering enough free features to attract a large user base while reserving premium features that justify a subscription fee. When a product has intrinsic value, such as Strava’s detailed performance metrics and analytics, the free version should provide sufficient value to entice users, but also create a natural progression to premium features for more advanced needs.

The researchers develop a dynamic framework to investigate consumer behavior in the freemium business model. Firms often use the free product for customer acquisition and introduce a referral strategy in which each successful referral provides an augmented free product to the consumer. This dynamic, while very helpful from a growth perspective, can potentially cannibalize the premium product. The authors use data from a storage and synchronization service to estimate a model of consumer behavior.

They find that consumers are forward-looking (i.e., they anticipate how they’ll behave in the future) and typically adopt the free product and make upgrade decisions in advance of reaching the baseline quota of the free product. Consumers refer friends, which can result in future increases in augmented baseline quota from a referral bonus when their friends join the service.

In addition, consumers actively manage their usage by deleting files to have an option value for keeping free storage space available on their accounts. They find that deletion and referral actions are often taken in conjunction when a consumer doesn’t wish to upgrade but wants to obtain more storage space.

In other words, the freemium version is effective in encouraging growth, but this growth may come at the expense of monetization.

The authors’ model relies heavily on the virality of the product.

Virality can exist in its own right as a growth strategy, but can also be due to network effects.

Just to be clear, a product can be viral without any network effects. Think about Survey Monkey, a highly viral product with no network effects. If more people use Survey Monkey… my life (or anyone else’s for that matter) won’t change much.

Role of Network Effects

And indeed people have studied the impact of network effects on the viability of a freemium business model.

Samuel Sato’s paper “Freemium as Optimal Menu Pricing” shows that the freemium model can be highly effective when the value of the network (and thus the willingness to pay) increases with the number of users. This is particularly true when network effects are strong. His paper delves into the freemium business model, particularly within the context of advertising platforms in two-sided markets. Both Strava and Runalayze make money from ads, but for Strava, this accounts for less than 10% of their revenue. However, I think the main rationale is still relevant. Many people are on Strava not only for tracking their runs, but also to see how their friends are doing and get motivated to run more consistently (note that I disregard any other use beyond running on Strava... Sorry dear cyclists).

The paper “Freemium Business Models as the Foundation for Growing an E-business Venture: A Multiple Case Study of Industry Leaders” by Günzel-Jensen and Holm discusses how companies can manage the lifecycle of conversions from free to paid users. The study emphasizes the role of free users in building a network and providing feedback, which allow the firm to tailor its business model to its customer needs. It suggests that successful freemium models continuously adapt their feature offerings based on user behavior and feedback. In other words, the existence of network effects at the free version level, together with a dynamic experimentation of features, allows firms sufficient time and base to develop the “right” business model, primarily in new areas.

Profitability and Growth Trade-offs

So far, we’ve dealt with the question of “when is it smart to offer a freemium model?” The question remains on which features to include.

The paper “Freemium Pricing: Evidence from a Large-scale Field Experiment” by Runge et al. suggests that reducing free features can increase conversion rates, but may also impact overall usage and engagement.

The paper aims to address the complex trade-offs involved in the design of freemium business models, particularly focusing on the impact of reducing free product features.

The central question of the study is: How do different configurations of free product features affect user engagement, conversion rates, and viral activity in a freemium model? The paper specifically examines the effects of reducing the number of free levels in a gaming app and the implications on user behavior and profitability.

The research employs an experimental approach, conducting controlled experiments to compare different configurations of free product features. In the game the authors’ study, after clearing level 40, players encounter a ‘gate’ that needs to be unlocked to access additional levels.

The study involves three experimental conditions

• A default condition with 40 free levels.

• The first treatment introduces an additional gate after 20 levels —players can only play 20 levels for free before encountering a paywall.

• The second treatment exposes players to the first gate after level 40 as in the default scenario, but the gate can no longer be unlocked with stars (by replaying levels prior to the gate).

The key metrics analyzed include usage rates, conversion to paid users, and viral activity (measured by referral rates).

The experiments reveal several key findings:

Treatment 1 (reduction to 20 free levels) showed that usage and conversion remained largely unaffected, but viral activity increased significantly by 28%. This suggests that while the core engagement might not change, users become more active in referring others when free features are reduced.

Treatment 2 resulted in a 7.9% decrease in usage but a substantial increase in both conversion rates (20.98%) and viral activity (15.81%). This indicates that more restrictive free features can drive higher conversions and referrals, albeit at the cost of reduced overall engagement.

The insights from this study have direct implications for Strava and Runalyze as they navigate their freemium models: While the reduction in free features can decrease overall usage, the significant increase in conversion suggests that this trade-off might be worthwhile.

It’s clear that both Strava and Runalyze need to identify the optimal balance where the reduction in free features maximizes conversion to paid users without excessively deterring usage.

At this point, Strava seems to be doing a good job… or maybe not. Current estimates show that only 2% of Strava’s users are paying customers, which is really small. 

Network Effects and Intrinsic Value in Strava

Note that all of these studies make one crucial assumption, that the basic product has network effects and/or is viral. So not only should Strava offer it for free, but the value should increase over time, either to the users (network effects) or to the firm (through virality).

Let’s try to understand whether Strava actually has network effects and exhibits viral growth.

I’ll use two different methods for this. The Bass diffusion model to estimate whether Strava’s growth indicates any virality, and Metcalfe's Law to identify actual network effects (in terms of increasing value to current and prospective users).

Bass Diffusion Model: 

The Bass Diffusion Model helps us understand how new users adopt a service over time, considering both independent adoption and social influence. For Strava, the model indicates significant virality with social influence playing a crucial role in user growth.

Estimating the model, we get:*<

• Coefficient of Innovation (p): 0.00807

• Coefficient of Imitation (q): 0.0788

• Market Potential (m): 1319.88 million users

And graphically:

The higher coefficient of imitation (q) compared to innovation (p) suggests that existing users significantly influence new user adoption, validating the presence of strong virality.

Let’s see if what drives this virality is indeed network effects.

Metcalfe’s Law

As we’ve seen before, Metcalfe’s Law posits that the value of a network is proportional to the square of the number of its users, as this is the number of connections between network members. Applying this to Strava, the value (measured through revenue per user) increases quadratically with the user base.

Bottom Line

The findings from the Bass Diffusion Model and Metcalfe’s Law highlight the effectiveness of both virality and network effects as the basis of Strava’s freemium strategy.

By offering essential features for free and leveraging network effects, Strava increases user acquisition and engagement. The social aspect of Strava enhances the platform’s value as more users join. Features that promote social interactions, such as segment competitions and social networking, should be emphasized to maximize network effects.

The question then is whether Strava just functions as a social network (which is already included in its free version) and the rest of its features are of no interest to the vast majority of its users. Users only want to see and be seen, and running is just one way of doing that.

But maybe this is very much 2024, where running is more of a social activity than a health activity. Or maybe we should understand that nothing is healthier than being social.