Traditional Product lifecycle identifies four phases of Product development: Inception, Growth, Stabilization, and Decline. However, an alternative can be proposed. We believe that initially, the team will add features to launch and grow the product like there is no tomorrow (The product's "Big Bang"), then comes a phase of technical improvements ("Future-proofing"), and finally, a moment where the founder can contemplate removing features ("Maturity").
In today's article, we're exploring when to remove features and how to best perform this process. For some of the founders, this phase might be far away, but eventually, the product will be stable enough that the user experience will benefit not from another risky experiment but from simplifying the product. Let's jump right to it and pose the fundamental question:
To add to the intro below, the phases listed are not mutually exclusive. There is nothing wrong in a situation where the founder issues new functionalities while removing old ones. However, in the initial fight for relevance and profit, there probably won't be any time to dedicate resources to removing anything unless it is vital to the product's survival. Only once the product's winning formula is found can the rest of the product align to fit the equation. Is there a sense in removing anything in the first place, as work done might benefit someone? Well, "The whole is better than the sum of its parts." Thus, when crafting a more final, user-friendly, easy, and responsive version of your product, you may find that some of the puzzle elements no longer fit. Thus, let's explore different scenarios when removing a feature is justified, starting with:
If data shows that customers rarely use a feature, then it's worth considering axing it. There are two reasons to do this in this scenario: Simplifying the interface and making it easier to find more popular options. But simplification is also beneficial under the hood! If the product's code has one less dependency and QA engineers have one fewer feature to consider when testing a new release, that saves countless development hours for more productive work. Since users don't use the soon-to-be-remove feature, it does sound like a win-win scenario. The best example here is the infamous Google Graveyard, with dozens of products killed due to low usage.
Now, all the arguments from the previous case still apply and have one more benefit: Improving overall user experience. Usually, this happens after a failed A/B test, but also, a once-popular feature may simply age poorly and, once beloved, may become hated. A scenario probably alien to any founder is when a feature is released, kept, and hated because of the insistence of corporate overlords. In that case, the feature is removed the moment a new management comes in with a different vision. A great example here would be Skype and its "Stories" feature mimicking the same functionality in Instagram. Not only did it never fit the communicator in the first place, but users hated it.
The technology develops every year and functionalities that used to be manual can now be easily automated. So, if you can solve a certain user problem automatically, without any user input, why bother with a dedicated feature in the first place? An example of a feature being removed due to advancements in automation is the Timeline feature in Windows 10. Microsoft axed this feature in Windows 11, as similar functionality became available through Microsoft Edge and other more automated solutions.
Now, there may be a feature that is used, even beloved, but when it hurts product responsiveness or the bottom line, sometimes there will be no choice but to remove it. We predict that unless Large Language Models become cheaper to run, shortly we will see an AI backstep era, where lots of chatbot support will be removed from products due to high cost. Another great example is when Reddit introduced changes to its API. This was beneficial for Reddit in two ways: it managed to kill several free competitors that used its content but didn't provide Reddit-associated ad revenue. It also lowered the cost of running the API service, as it was no longer utilized as much. Granted, the API wasn't removed per se, but it's the next best thing that will come back to in this article.
Sometimes products and features will be killed as the company changes strategy, vision or simply decides to focus on one thing and avoid distractions. An example here would be the "Trending Topics" feature on Facebook. It was initially designed to help users discover popular conversations across the platform, but it was eventually removed. This change was part of a broader shift in Facebook's strategy to promote community engagement and trustworthy news over viral content.
In some cases, changes in laws or industry standards can make a feature non-compliant, and removing it is the only way. A lot of video games had to remove "gambling-like" features when the regulators started pushing back on such an exploitation. Of course, this is a drastic example of a feature that would have come to be in an ideal world, but changing regulations can also affect less sinister elements of the product.
Now, some features may have been added with a very poor technical or user interface implementation. Rather than fixing or updating it, sometimes it is best to leave it behind and again, save countless development hours. Nothing is stopping the founder from reinstating the feature in a more state-of-the-art state, that will no longer jeopardize the product's technical health.
This is a consideration that is growing more prevalent as time goes by, and both security and privacy are in the spotlight of any product. The cases where products had to remove features to protect their users were more akin to earlier Internet. A great example would be the situation where a system could tell a user if a certain email was registered or not on a specific site using a password recovery tool. While it doesn't sound as bad at first, it poses a significant danger to individuals in the context of unsafe sites. Journalists could probe the sites to uncover popular people using the site, while external bad actors could even register a popular figure on the platform to slander someone.