Breaking Down Audible’s Churn Challenge
How to reduce user bottlenecks when selecting new audiobooks
Thanks for tapping in and reading. If you enjoy this piece, consider subscribing and sharing with your friends.
Today's article focuses on the most critical factor Audible needs to master to retain users—helping them discover new audiobooks to listen to.
Most people who purchase a subscription to Audible do so to continually listen and find new content. They aren’t one-off buyers of an audiobook. However, finding a new audiobook to listen to can be frustrating.
The Pain of Deciding on a New Audiobook to Buy
For all Audible users reading this, think about when you are about to finish a wonderful audiobook. You see that a new credit will hit your account in five days. You quietly ponder which new audiobook you want to listen to.
You start scrolling through your wish list but aren’t sure what to select. You listen to a sample of an audiobook that looks interesting, but the narrator sounds a little monotone. You find another audiobook, but a well-written review mentions the slow buildup.
Four days pass and instead of deciding on an audiobook to buy, you cancel your monthly subscription until you figure out what you want to listen to next.
As a regular Audible user myself, I have experienced this pain point. Canceling my subscription is something that I’ve regularly done when I can’t make up my mind on what to listen to next. It’s my biggest pain point and, I suspect, a major cause of churn.
If this is unfamiliar to you, let me give a quick overview of how Audible’s business model works. Most users purchase a monthly subscription plan, which delivers one credit per month. With that credit, you can select an audiobook to listen to. Every month, you receive another credit.
If you fall behind in your listening, you'll accumulate credits. However, if you cancel your subscription before using your credits, you lose them. Users are incentivized to use their credits before canceling, or in my case, cancel before receiving a new credit.
Business Goal
Now that we’ve identified the user pain point, let’s focus on the business goal. Audible makes money by keeping people subscribed and earning credits. For a user who is already converted into a customer, the goal is to reduce churn.
To measure success, I would track two key metrics:
New audiobook Starts: This indicates users are finding new audiobooks to start
Daily Listens: This reflects users getting value from the audiobook and are more likely to stick around
At first glance, New Audiobook Starts may seem counterintuitive. It indicates users are using their credits, which makes it easier to cancel a subscription if they can’t find something they want. But hold tight, I have a potential solution to fix the leak in the incentive structure—stay tuned!
On the other hand, Daily Listens are intuitive. It means that a user is enjoying their time on the platform, which is a strong indicator that they will stick around. However, daily listening only lasts as long as there are pages left in the audiobook. Loyalty is only shown to the audiobook; not to Audible. That’s why New Audiobook Starts are so important. It’s the metric that helps keep the cycle going.
Audible’s current experience is built to improve new audiobook starts and promote daily listening by minimizing pain points when selecting a new audiobook (spoiler: I think they can be doing better though!). Let’s look at what Audible is doing to drive these metrics and solve users’ pain points. I’ll share some of my personal thoughts along the way.
Recommendation System
Audible’s recommendation system suggests audiobooks similar to what users have listened to, helping them discover other titles they might enjoy. This is likely an effective way to get readers to continue to use their credits on genres they are interested in. However, in my experience, the recommendations can be repetitive and fail to spark new discoveries, becoming too one-dimensional.
Consider TikTok’s algorithm. It serves users content they dwell on and engage with, but it will also serve new content to see if users have other interests they want to consume. Audible would likely benefit from employing a similar strategy.
Audible could use signals such as previous audiobooks the user has completed (especially if it was a title read daily). They could also use the data of similar listeners who are listening to the same audiobook or the same genre to recommend titles from their listening history.
Samples & Reviews
Surely, one of the best selling points for users before purchasing a new title is to listen to a couple minutes of an audiobook's sample. In audiobooks, the narrator is just as important as the content of the audiobook itself. If you read through the reviews of any given audiobook, you are almost always going to see some critique of the narrator.
There is no easy solution to this problem, but with the development of Generative AI, there could be opportunities to create more engaging narrations. Perhaps it’s an extra offering that others can pay for (it would probably be cheaper than paying a human narrator), or maybe it’s something Audible can eat the costs for to improve engagement with their wide selection of audiobooks.
Similarly, it would be easy to sort reviews by star ratings and bury the poor reviews unless a user seeks them out. This is a bit of a dark pattern though, so I wouldn't necessarily recommend it.
Wish List Building
One of the best ways Audible helps users find and remember interesting audiobooks is by allowing users to build a wish list. It is generally the first place I go when selecting a new audiobook. However, I imagine many users are like me and end up having a wish list that has 50+ audiobooks on it. By the time a user has over 10 audiobooks on their wish list, my guess is that choosing an audiobook becomes more difficult and takes longer, increasing the risk of a user deciding to cancel their subscription.
If I were Audible, I’d make the wish list dynamic, reorganizing titles based on a user’s recent searches, engagement, and listening history. Similar to the recommendation system, I'd also prioritize titles that other similar users have enjoyed. These changes would help reduce decision paralysis and the endless loop of listening to samples and reading reviews.
All Out Audiobook Attack
After a user completes an audiobook, the only thing Audible should do is go on a full-out attack to get users to select their next audiobook. The long, drawn-out process of selecting a new audiobook needs to be simplified and personalized as I've outlined above.
Moreover, Audible should aim to reduce the perceived commitment when choosing a new audiobook. One experiment Audible could try is allowing users to reverse a credit within the first 30 minutes of listening to a new audiobook.
The last thing Audible wants is for a user to buy a new audiobook and quickly find out they are disinterested. This is the worst-case scenario and is likely a high driver of churn.
Similarly, reversing a credit and building up a bank of credits isn’t a bad thing because users are less likely to cancel a subscription due to a loss of credits while doing so. An idea like this helps people find audiobooks they enjoy so Audible can turn users into Daily Listeners.
Learnings
Unblock Bottlenecks: When users feel friction, their first instinct is to churn. Focus efforts on unblocking bottlenecks in the user journey to patch up a leaky bucket.
First Impressions Matter: Build personalized recommendation systems for individual wants and needs. Audible’s ability to keep users engaged and reduce churn depends on it.
Quality Over Equity: Recommend high-quality content over poor-quality content. Don’t be equitable in your recommendations. In the case of Audible, recommending audiobooks with great narrators will prevail.
Provide Flexibility: Allow users to test and try your product to get them hooked. Fighting against flexibility when it’s in the interest of the user and the long-term interest of the business is death by a thousand paper cuts.
Did you like this breakdown of how Audible tailors their experience to drive core metrics forward? If so, you can support me by subscribing and sharing this with your friends.