Startup Underdogs Ed. 7 - Managing our B2C Subscription Churn

It’s never fun to see customers unsubscribing from your product. Beyond the obvious cash flow and valuation implications at its core, it’s people telling you that they aren’t seeing the value for the price they’re paying.

In an ideal world, you have a hyper-focused ICP, a product that delivers immediate, tangible value, and a customer base so loyal that they say “hell yeah!” every time they use it.

But more often than not, as SaaS founders, we launch quickly, star acquiring customers, and suddenly find ourselves sweating bullets while staring at a daunting churn number.

Painful beginnings

When we launched our B2C subscription product, we were riding high on the excitement of early success.

$16,000 of new recurring revenue, in a single month.

Then, $62,000 of new recurring revenue.

This is incredible. Every month, we'd run the numbers and tell ourselves we simply needed to convert a small percentage of our 1,500,000+ users and we are rolling in cash.

Three months in, however, the cruel reality of subscription revenue set in. While we were starting to stack new subscribers month-over-month, we were also seeing subscribers unsubscribe. Ugh...

Churn...sigh...

I suppose it was to be suspected. We were supporting a community of 1,500,000+ people from all different countries, backgrounds, years of experience, intentions. While the flagship offering of our product (Career Paths) was launched with the intention of serving a targeted ICP, we were pursuing a bigger prize and we extended the subscription product offering to offer so much more to so many more.

Churn was pretty crazy the first few months, jumping as high as nearly 25% five months in. It was certainly a leaky bucket, with revenue falling out as quickly as it was coming in.

It was a major problem. And it needed to be solved.

Tackling Churn

Churn is never a "solved" problem. You are simply looking to get it to a manageable state that support positive economics in your business.

For us, we were advised to get churn down between 7-10%. And that became our target.

We knew quite a bit about our learners, who they were, how they thought, and what they valued. But we knew this as generalized knowledge, we couldn't point to discrete data points to slice and dice in reports.

So that's where we started - we sought to learn more about our learners.

The first form of this was an information capture as part of our onboarding experience.

We wanted to know more about someone not only to better market and communicate to them. We also wanted a better picture of the 60,000+ new registered users who were joining each month.

Which segments were more engaged with our free product. Which segments were upgrading to our consumer subscription, which had the highest propensity to churn, which had the highest LTV, etc.

(Oh, and who might be useful for our sales team to reach out to...??)

The first version of the capture, we asked:

• Role
• Company
• Years of Experience
• Goal (Get started in cybersecurity, prepare for an industry certification, enroll in a career path to get a job in cybersecurity, learn new skills, train my team)

At the time, we were seeing roughly 2,000 new registered users every day with 150,000-200,000 MAUs - we were able to gather a significant data set quickly.

But it only left us with more questions.

So we went deeper. For each of these "goals", we were these people? What did they care about?

• Cybersecurity Career Starters: These were individuals we either were in IT and wanted to move into cybersecurity; or, they heard about cybersecurity being a growing, lucrative career field and they wanted to see if it was for them. They value structure (can't give too much too quickly, can't overwhelm), simplicity, and seek our mentorship and guidance.
• Certification Seekers: These were individuals at all stages in their career. They could be just starting out looking for an entry-level certification (A+, Network+, Security+), or they could be experienced practitioners and managers (CISSP, CASP+, CISM, CISA). The in-person certification market was dominated by pass rates and guarantees; online, it was all about building confidence towards passing your test the first time.
• Job-Seekers: This was a complicated segment, as it was made up of opportunists (people who sought to get a job to make great pay) as well as genuine, highly motivated people who are made to be great cybersecurity professionals. Regardless, this segment often sought out guarantees or program credibility to get a job.
• Skill-Enhancers: We knew this group existed, but it was the most challenging group to navigate. The people purely there to learn new skills passively, to stay up-to-date with what's moving in the industry. Because the intent isn't necessarily tied to dollars and cents, it is a fickle group requiring constant new content to please, while toiling against low engagement and price sensitivity.

These were helpful, as we built marketing campaigns and communication cadences to help us acquire, engage, and position our product to impact each group.

As a result we started to see churn come down slightly.

"The hardest part of managing product is sifting through and identifying the signal from the noise."

Over the next few months, as we gathered more data, it felt like we were still missing something critical to way these learners ticked.

The answer to that question came from a conversation we had with a highly engaged user who also happened to be a gym fanatic. They helped us understand the world of online learning by comparing it to gym goers.

In hindsight, it made a TON of sense.

In both situations, you typically have this overarching goal you are "working" towards. And, unfortunately, to reach that goal, it takes time and consistency; as well as, forcing yourself to be uncomfortable over and over again. You'll feel lost, feel like you're going backwards, feel like it's not working.

While we never incorporated it into our information capture (would have been interesting if we did though!), we started thinking about our learners in the following buckets:

• Self-motivated ("lone gym goers"): These were people who could tie the value of the product to their own personal goals without needing external triggers. For our learners, they saw how they could use our Certification Prep programs or Career Paths to get themselves certified, a new job, or promoted in a cost affordable way. The resources were there for them to succeed, given a little intrinsic motivation.
• Coach-dependent ("seek trainers"): These people need others to hold them accountable. They crave clear pathways, direct mentorship, and on-going nudges to keep them on track. They can fall off quickly without accountability. It feels like a slog to do the work, but they continue showing up given the proper structure is in place.
• Program-focused ("class enthusiasts"): These people crave community connection and accountability. In the online world, these are known as cohort-based learning programs. The presence of the instructor (or mentor) and peers provides them what they need to show up, do the work, and receive the results.
• Casual learners ("drop-in gym goers"): These people show up without a goal or intention. They are happy to learn new skills, or keep up with something new or interesting. They are happy to dabble in a variety of different things but they lack that greater goal. And without that goal, are vulnerable shiny object syndrome as new competitors emerge.

Building a Success Story

To drive growth in your company or your product, a simple model is:

• Do people know you?
• Do people like you?
• Do people trust you?

We knew people knew us and we knew people liked us. But this was a new product we were promoting. This wasn't our free training. They were putting down real dollars to solve a real problem they had, and we weren't convinced people trusted it could do it.

We believed in our product (you have to). We believed it could impact tends of thousands, if not hundreds of thousands, of people. Cybrary was all over Reddit, Twitter, Facebook - people were recommending us to anyone who wanted to get into cybersecurity or needed cybersecurity training.

While this was great, we wanted create assets we could use to build trust. Something tangible. Something we could point to and say "this works".

We asked and the universe provided - not 6 months from launch, we were poking around our user community and discovered one of our learners went from zero to cybersecurity in three months.

As marketers, we were giddy at the prospect of turning this story from a half-cocked conversation into full blown PR and marketing campaigns.

What a win!

https://www.bizjournals.com/washington/inno/stories/profiles/2018/10/04/how-marylands-cybrary-turned-a-mechanic-into-a.html

https://www.ciodive.com/news/online-cybersecurity-training-propels-students-into-industry-jobs/539486/

https://www.youtube.com/watch?v=_I1jLEnodzs

But wait, there's more...

A great success story wasn't the only lever we pulled. Yes, it was incredibly helpful, using it in marketing campaigns in attempts to attract more ideal customers (and in turn lead to better retention); but battling churn requires a small, nimble team testing and iterating over solutions.

Post-Purchase Flows

Around September 2018, it was becoming painfully obvious churn was totally out of control. Much like any problem, mapping the customer lifecycle (including communication touch-points) became a critical tool for us.

We had been spending so much time looking at improving things pre-purchase, we completely forgot about what was happening post-purchase.

Once a customer purchased, we'd drop them back into the product, and that was it! ??

Oops!

So we went to work experimenting with onboarding guides, post-purchase screens, etc. We crafted post-purchase onboarding email cadences helping them learn not only how to use the product, but how others were using and finding success from it.

Paid Slack Community (Mentors and Peers)

Our paid Slack Community was one of the most impactful engagement tools for our subscribers. Unfortunately, as we got north of 17,000 subscribers, we never quite cracked how to leverage Slack to support learners' needs.

Early on, however, we found that someone who participated in the community was orders of magnitude more likely to retain. So this became an important product usage metric for us to track, and we were intentional on not only making sure people were aware it was available to them, but actively pushed it on people (sometimes you need to be a bit annoying to get people to do something that's in their best interest ??)

Cancelled - Not Paid

By far the most frustrating part of fighting churn was the fact on any given month, 30% of subscribers who churned, did so with the status Cancelled - Not Paid. Granted, if those learners were active and engaged with the product, they would have never let their subscriptions go unpaid. But the dirty little secret of many online subscriptions businesses, is they are more than happy to have you continue to pay, even if you don't use the product.

We ended up becoming early customers of a company called Gravy Solutions, a payment recovery solution. They would work inside your subscription management platform to optimize your Dunning Management and/or extend your payment recovery time period with manual outreach to subscribers. While Dunning Management was the most effective way for us bring down churn a few percentage points, Gravy certainly helped some - enough to cover their bill.

Annual Plans

I'd be remiss to say our work on marketing, product, and billing management were the keys to bringing down churn like it did.

The thing which brought the monthly number down to the 7-10% mark was a more palatable price point and a intentional move to push toward annual plans.

"Hey now!" says the observant reader.

"If you are selling annual subscriptions instead of monthly subscriptions, who's to say you impacted churn; rather you delayed the problem and made it a future you issue!" - Observant Reader

You're 100% right in that comment too. But we were looking to buy time. Time to work on the product, reduce friction, and to get our flows right. Time for learners to experience the immense value they could receive from our product (remember, slow time-to-value!).

I wish I still had access to show the performance of mature cohorts, but we did find a noticeable improvement.

Annual cohorts had about 50% retention after Year 1, then 25% retention after Year 2.

Monthly cohorts had about 50% retention after Month 3, and were down around 10% retention by Month 12.

^ The price of each obviously matters when calculating LTV, but prices fluctuated so much, you'll have to take my word when I say Annuals performed better - more cash upfront, and better LTVs. ??

As a result of this, our average price point decreased significantly (we were selling more annual plans at a lower per month price point than the monthly plans).

Final Takeaways

The B2C Product Line peaked around $7.5million ARR in 2020 during peak COVID. Over the next couple years, B2C churn would creep back up to the 10-15% mark. For a variety of reasons, we could never reel in the churn number. There were many contributing factors which aren't necessarily appropriate to write about. At the end of the day, high churn is a product-market fit problem. As new and legacy competitors entered the market with better products and attractive price points, we found ourselves sitting on a stale product, having been out-innovated and out-maneuvered.