Strava recently decided to raise its prices in the worst possible way—not telling anyone and letting the internet figure things out on its own. Wildly diverging price increase figures began rolling in from different places, leaving angry nerds to do what they do, and letting the meme pages feast on the scraps.
Nice Of You To Get Back To Us
Apparently tech publication The Verge is a big enough deal that Strava felt compelled to make an actual reply:
“There are a lot of learnings here, but we chose to do it this way for very specific reasons,” Strava spokesperson Brian Bell tells The Verge. “We thought it was best for our business.” Bell says the company decided to alert members on a rolling basis because they chose to subscribe at different points throughout the years.
The company’s choice to claim the objectively terrible rollout was done on purpose is certainly something. But buried deeper in the article is a very alarming tidbit (emphasis mine).
Some of the confusion, he says, is because customers are seeing different rates of increase based on when they joined. The most impacted group, Bell says, will be the 30,000 users who started subscriptions before 2016. However, everyone affected will end up paying the same new price based on their region.
As someone who’s been digging into Strava’s user numbers on-and-off for the past decade, this 30,000 pre-2016 subscriptions number is alarmingly small.
A History of Paying Pre-2016 Strava Users
In 2013, Strava told Outside that 20% of its users paid for the service, but declined to specify how many total users they had at that time. But because Strava assigns ids to new users with consecutive numbers, by looking when users in a given set of ids began uploading activities, we can infer total users in Jan 2013 were around 1.5 million, giving us a total of around 300,000 paying subscribers.
In 2015, Strava user and friend-of-Cyclocosm Mark Slavonia used a similar approach to arrive at a figure of 192,000 paying users from 8.2 million in total. That would seem to be a decline, but his sample size (384) was relatively small, and his “95% confidence” window returned between 65,000 and 311,600 paying users.
I ran my first automated sample of 5000 Strava users in 2017, and found that of the 2386 created pre-2016, 81 were premium members, or 3.3% The last pre-2016 user in that sample created his account on 31 Dec 2015, and had a user a user ID of 12657818. Multiplying that number by the premium rate gets us 429,707 paying pre-2016 subscribers.
After an Outside article repeated Mark’s 2015 analysis and found a paying subscriber rate 1.5-2 times what I had seen previously, I ran another automated 5000 user sample (it revealed at 3.16% subscriber rate, still well under Outside’s figures). While changes to Strava’s website removed the user creation date from what I could access, users ids below 12657818 (and thus created before December 31st, 2015) were paying at a rate of 4.91%, inferring 530,362 paying pre-2016 subscribers.
Shortly after summing up my festering Strava complaints for OneZero in 2020, I ran another random 5000 user sample, my most recent. Of the 927 users with IDs below 12657818, 57 were premium members—a startling 6.15%. Extrapolating against the 12,657,818 users created before 1 Jan 2016, that’s 778,455 paying pre-2016 subscribers.
Literal Order of Magnitude Decline
First a caveat that the above numbers are rough. While I’m fairly confident in the technical accuracy of my data gathering, I’m nobody’s StatsBoi and my assumption is I’m probably doing something wrong somewhere.
Plus the share of pre-2016 accounts declined each new time I took a sample (because I pulled randomly from the pool of overall users) making each batch less accurate for pre-2016 accounts than the one prior. And of course, all I can see from the website is what’s on the website. I have no idea whether a premium status is promotional, how long its been held for, how much it cost the user, etc.
But even accounting for all of that, a decline in paid, pre-2016 memberships to just 30,000 would be catastrophic. Compared to every estimate above, it’s an order of magnitude decline—like literally, to the point that I reached out to The Verge to see if it might be a typo for 300,000. (Update: The Verge has confirmed it is not.)
It’s also possible that Strava meant this “most impacted group” was just the 30,000 pre-2016 accounts that had been billed at the new rate(s) so far? Or could it have been only pre-2016 accounts that had never changed their subscriptions? The lack of a direct quotation containing the figure further muddies the water. Certainly some kind of follow-up is in order.
All About That (User) Base
Stepping back for a moment, it’s kind of amazing to look at that historical sampling data and see how statistically loyal Strava’s earlier adopters have been to the platform—at least prior to yesterday’s number.
Even through the company’s much-maligned pseudo-pivots, privacy issues that are still exposing sensitive infrastructure, and relatively limited response to consumer feature requests, these users have continued shelling out fees, keeping the lights on as Strava struggles to justify its colossal valuation.
What’s clear from this (latest) kerfuffle that is Strava (still) needs to treat these core users better. The lack of any heads-up, the delayed and ambiguous feedback, the inability to tell the people who are actually paying what they’re actually paying—all reveal a pretty stark disconnect between the core value proposition of the product and the people running the company (sound familiar?).
My fear is that years of dogged user loyalty may have created an impression at Strava that the company can do whatever it wants with little fall-off from the paying membership. If this 30,000 pre-2016 subscriptions number is accurate, that assumption may have proved to be cataclysmically incorrect.