"As a result, our future growth and financial performance may depend, in part, on our ability to sell more subscriptions to our premium services," Fitbit (FIT) wrote in its original S-1 Registration Statement back in 2015. The context was the risk associated with relying predominantly on hardware sales, while the company hoped to grow its subscription business. Here's the full quote:

However, in the future we expect to increase sales of subscriptions to these services. Our inability to successfully sell and market our premium services could deprive us of a potentially significant source of revenue in the future. In addition, sales of our premium services may lead to additional sales of our connected health and fitness devices and user engagement with our platform. As a result, our future growth and financial performance may depend, in part, on our ability to sell more subscriptions to our premium services. 

For a while, this risk wasn't all that pressing, as unit sales and revenue continued to march higher following the IPO. With sales on the rise, the desire to focus on subscription-based revenue could wait. But then, the wearables market started to shift away from basic trackers and toward smartwatches.

3 Fitbit Ionics on top of each other

Image source: Fitbit.

Devices and users aren't doing so hot

The timing couldn't have been worse. Fitbit went public just a couple months after Apple launched Apple Watch in 2015, its first smartwatch that undoubtedly catalyzed the broader market's transition away from fitness trackers. Fitbit was able to enjoy a strong holiday shopping season that year, peaking with 8.3 million devices sold. Unit volumes have been heading the wrong way ever since.

Chart showing Fitbit's devices sold declining recently

Data source: SEC filings. Chart by author.

Beyond just devices sold, Fitbit also reports active users on an annual basis. There are troubling signs here, too. Last month, Fitbit announced it had reached 25 million active users by the end of 2017. That's a respectable milestone, but it also represents significant deceleration from prior years. Put another way, Fitbit added fewer than 2 million active users last year, after adding 6.3 million in 2016 and 10.2 million in 2015.

Chart showing user growth slowing

Data source: SEC filings. Chart by author.

It goes without saying that devices sold are effectively the pipeline for active users, but the majority of Fitbit customers don't remain active on the platform. At the end of 2016, when Fitbit had 23.2 million active users, it had 50.2 million registered users -- over half of those registered users had become inactive.

Fitbit is quick to point out that users can become active again, and advises investors not to read too deeply into the discrepancy between active users and registered users: "Therefore, you should not rely on our registered device user metric as an indicator of the level of retention of individual users in the future, continual user engagement, future payment by users, nor the potential size and growth of our user community as an indicator for other revenue opportunities, such as subscription-based premium services and our corporate wellness offerings."

Still, the data speaks for itself.

Neither are subscriptions (yet)

Subscription revenue has historically accounted for less than 1% of total revenue, and Fitbit has never been able to grow that business beyond that threshold, though not for a lack of trying. The company had acquired Fitstar shortly before the IPO to add a subscription-based fitness service to the mix. The company also used to offer a Fitbit Premium service for $50 per year. Both of these have been consolidated and rebranded under Fitbit Coach, which launched alongside Ionic last year for $40 per year.

Fitbit reports fourth-quarter earnings on Feb. 26, but we know total revenue in the third quarter was $393 million. That means subscription revenue was $3.9 million or less. Fitbit Coach launched in the fourth quarter, so Fitbit Premium would have been the subscription service available in the third quarter. At $50 per year, Fitbit would only recognize $12.50 per quarter per user, with the rest being deferred over time. That implies that Fitbit had at most 312,000 paid subscriptions in the third quarter -- a minuscule proportion of its active user base.

Furthermore, deferred revenue has been declining, and stood at $40 million at the end of the third quarter, down from $49.9 million at the end of 2016. That trend shows that as Fitbit recognizes deferred revenue over time, it's failing to build a pipeline of ongoing deferred revenue going forward. It's worth pointing out that Fitbit's deferred revenue does not consist entirely of subscriptions; corporate wellness software, software upgrades, and other free services that Fitbit offers are included in that line item. (Even free services and software upgrades must technically carry an accounting cost.)

Unit sales are declining while active user growth is slowing to a crawl. Fitbit desperately needs to build its subscription business.