Lenny's Newsletter · Product & Work
TIER 4 2022-09-13
> ## Q: I’m struggling to decide if I should sunset a feature that’s taking up a bunch of resources. While at Airbnb, I saw many products launched, and surprisingly many sunset. Including many that had launched with much fanfare, like [Guidebooks](https://www.pcmag.com/news/airbnb-app-gets-personalized-matching-guidebooks), [Wish Lists](https://techland.time.com/2012/06/27/airbnb-gets-more-social-aspirational-and-beautiful-with-wish-lists/), [Neighborhoods](https://techcrunch.com/2012/11/13/airbnb-launches-neighborhoods-providing-the-definitive-travel-guide-for-its-guests/), [Match](https://thenextweb.com/news/airbnb-launches-its-match-service-to-help-you-book-places-to-stay-in-a-pinch), [the Apple Watch app](https://www.theverge.com/2015/9/3/9250379/airbnb-launches-apple-watch-app), *[Pineapple](https://www.adweek.com/performance-marketing/airbnb-launching-print-magazine-161478/)* [magazine](https://www.adweek.com/performance-marketing/airbnb-launching-print-magazine-161478/), and even [Stories](https://www.cntraveler.com/story/airbnb-launches-instagram-like-stories-feature).  Looking back, not once do I recall that anyone regretted a product or feature being sunset. If anything, teams waited too long to kill these products. But that’s not always the case. When I [polled people on Twitter](https://twitter.com/lennysan/status/1409934916754710528), a few compelling counterexamples came up: Kill a useful feature prematurely and slow future growth—or spur the ire of your most loyal customers. Keep an unnecessary feature around for too long, and precious resources are wasted on low-impact work. Based on some research and conversations with colleagues, I’ve come up with a handy checklist to help decide when it’s time to sunset a feature. If the product or feature checks three of these five boxes, it’s probably time to shut it down: 1. **Low usage:** Less than 5% of your active users are engaging with it. 2. **High cost to maintain:** Takes more than 10% of your team’s resources just to maintain it. 3. **Degrades the user experience:** Gets in the way of users completing important tasks or adding important new features. 4. **Misalignment with strategy:** Does not support your strategy, and is very unlikely to align with your strategy a year from now. 5. **Very few vocal or important users:** You aren’t worried about an outcry or a big revenue hit. Here’s a bit more on each of these points. #### 1. Low usage Obviously if no one uses the product, it’s a great candidate for being sunset. What percentage is low enough? Five percent is a good starting point, but I encourage you to think about it this way: if you were to launch a new product/feature, what percentage of adoption would you consider success? Any feature with lower adoption than that baseline is worth paying attention to. Note, though, usage can sometimes be deceiving, in that a customer may use a feature passively but be totally OK with it going away. As TK suggested below, another way to track usage is how many people complain about it when it isn’t working. *Key question: What features are used by less than 5% of your users?* #### 2. High cost to maintain The idea to sunset a feature generally surfaces from an engineer or a PM whose team is responsible for maintaining a barely used feature. It’ll feel like a huge waste of time—especially when it’s something that got dumped on your team. My team once spent about five months building a product called Milestones, which was designed to delight Airbnb hosts with in-product messages celebrating their hosting milestones (e.g. hitting their 100th review, having a guest from a faraway country, hitting a perfect response rate, etc.). I was really proud of the product, and there was a lot of confidence internally that it would meaningfully impact host engagement, retention, and satisfaction. It did none of these things. Zero measurable impact. After a reorg a few months later, another team ended up owning this product. Maintaining the product quickly began to suck up a large chunk of this team’s resources. Even just fielding bugs ended up taking a lot of the PM’s time. So in the next planning cycle, the PM proposed that this product be sunset. After much debate, leadership eventually agreed, and they were able to kill it and move on to much more important work. Looking back—great decision. How much maintenance time is enough to consider sunsetting a product or feature? I’ll suggest a threshold of 10% of your team’s resources. If 1 out of 10 engineers are taken up maintaining a legacy feature that no one is planning to invest in, it’s a great sunset candidate. *Key question: What legacy feature or product (that isn’t contributing to your team’s goals) is sucking up 10% or more of your team’s resources?* #### 3. Degrades the user experience Another sunset candidate is any feature that gets in the way of your user accomplishing important tasks. [Anu Atluru](https://anuatluru.com/), an early employee at Clubhouse, [shared](https://every.to/p/what-i-learned-at-clubhouse) a great takeaway from her time watching the Clubhouse product evolve: > *“The first version of your product is usually simple. As you keep building, the product becomes more robust in quality, utility, and experience. But there’s a tipping point after which ‘robust’ can sneakily switch to ‘complicated.’* > > *Some signs your product might be too complicated: You’ve launched a bunch of features and haven’t killed any of them. You’re running out of literal surface area to put new features. Users aren’t clear about what to do with the product.”* The [Airbnb Neighborhoods](https://techcrunch.com/2012/11/13/airbnb-launches-neighborhoods-providing-the-definitive-travel-guide-for-its-guests/) product is a good example of this. On the surface, giving people insight into the character of each neighborhood when they are picking a location to stay in sounds incredibly helpful. It turns out, though, people get distracted reading about all of the interesting neighborhoods and end up not booking an Airbnb at all. Though the product, and the content, was beloved, it increased friction for the most important user flow. If a feature is consistently keeping your users from accomplishing key tasks, or is making it significantly more difficult for you to design new and important functionality, it’s a great sunset candidate. *Key question: What feature is making it harder for your users, or your designers, to accomplish more important work?* #### 4. Misalignment with strategy A product may be rarely used, hard to maintain, and degrade the user experience, but if it supports your company [strategy](https://www.lennysnewsletter.com/i/1225083/what-is-a-strategy), then it can still be worth keeping around. Who knows, you may decide to invest in it in the future, or it may be important to your broader story to customers or press. If, however, the product or feature has nothing to do with your strategy and is just there because it’s there, it’s a great sunset candidate. The Milestones product I mentioned above is a good example of this. When it was built, one of our company strategy pillars was increasing host satisfaction. After taking a number of shots at it, including some meaningful wins like the [Superhost program](https://airbnb.com/superhost), we moved on to other priorities. The Milestones product was no longer aligned with our strategy, and as we saw above, was taking a lot of resources to maintain. Thus, it became easier to support the idea of sunsetting, in spite of all the work that went into it. The Superhost program, on the other hand, also saw little measurable impact initially and took a lot of work to maintain, but it was squarely aligned with our long-term strategy (i.e. treating Airbnb hosts like partners, and making it easier for guests to find amazing places). So that program continued, and has become a core part of Airbnb’s DNA. *Key question: What’s a legacy product you’re currently maintaining that no longer aligns with your company strategy?* #### 5. Very few vocal or important users Finally, an important consideration when sunsetting a feature is how much blowback you’ll experience if you sunset it. For example, here’s [Gibson’s](https://twitter.com/gibsonbiddle) experience at Netflix: Also, investigate how important this feature is to that small minority of users: On the flip side, [when Facebook introduced the news feed](https://www.wired.com/2016/09/everyone-hated-news-feed-then-it-became-facebooks-most-important-product/), there was mass freak-out for a while, and then everyone got over it. Now it’s the default for how we experience content on basically every platform. In my experience, if you’re confident a change needs to happen, talk to users who are still using it to see if you’re missing something, but generally index on the side of pushing through the blowback—make the change, sunset the feature, and refocus your efforts on making your core user experience faster, simpler, smoother. > #### “Folks love to build stuff. From time to time, though, you need to prune your product. Get rid of stuff that is no longer relevant. If you don’t, your product will quickly get filled with bloated complexity. In consumer software, keeping things simple is almost always better than solving everyone’s problems, for both the consumer and business.” > > #### —Gibson Biddle And finally, what comes after you’ve decided to sunset the product or feature? Here’s a great simple set of next steps from [Ryan J. Salva](https://twitter.com/ryanjsalva/status/1569714818147581952) (VP of Product at GitHub): If you’ve got any experience sunsetting a product or feature, good or bad, I’d love to hear it 👇 [Leave a comment](https://www.lennysnewsletter.com/p/when-to-sunset-a-feature/comments) Happy sunsetting!  ### 📚 Further study 1. [How to sunset a feature](https://www.intercom.com/blog/how-to-sunset-a-feature/) by Intercom 2. [The art of unshipping: How to deprioritize features, phase out projects and sunset products](https://plan.io/blog/deprioritize-and-sunset-projects/) by Planio 3. [Quick Guide: Sunsetting a Product](https://productcoalition.com/i-was-recently-reading-todd-berkowitzs-blog-overcoming-the-sunk-cost-fallacy-when-selling-907963fb5192) by Krishnan Hariharan *Have a fulfilling and productive week 🙏* ## 📣 Join Lenny’s Talent Collective 📣 If you’re hiring, [join Lenny’s Talent Collective](https://www.lennysjobs.com/talent/welcome) to start getting bi-monthly drops of world-class hand-curated product and growth people who are open to new opportunities.  If you’re looking for a new gig, join to get personalized opportunities from hand-selected companies. You can join publicly or anonymously, and leave anytime. [Apply Now](https://www.lennysjobs.com/talent/welcome) #### **🔥 Featured job openings** 1. **Slab:** [Frontend Engineer](https://www.lennysjobs.com/jobs/900e2226-f222-4108-90a8-285824523cec) (Remote) 2. **Walmart:** [Sr/Principal Product Manager (Store8 incubator)](https://www.lennysjobs.com/jobs/22be805a-9d53-46f9-ac69-ddf6b158c2fb) (SF, Manhattan, NY, San Diego, CA, Seattle, WA) ## **🧠 Inspiration for the week ahead** 1. **Listen:** [Tools to Improve Your Focus & Concentration](https://open.spotify.com/episode/0cfMpJd2lk9xtk3hyGy5FT?go=1&sp_cid=8e64ff43e771fbb17cb0525df83e47e4&nd=1), Huberman Lab 2. **Read:** [Lifestyles](https://collabfund.com/blog/lifestyles/) by Morgan Housel 3. **Read:** [Notes on Progress: An environmentalist gets lunch](https://worksinprogress.substack.com/p/notes-on-progress-an-environmentalist) by Works in Progress **If you’re finding this newsletter valuable, consider sharing it with friends, or subscribing if you haven’t already.** Sincerely, Lenny 👋