Removing features from a product is like decluttering your home. Everything in your home was important to you at some point in time; you bought it with your hard-earned money, you put thought and effort into it, and you simply don’t want to get rid of it. Yet, that which no longer serves you must go.
Just as you don’t want a cluttered home, you don’t want a product cluttered with way too many features. Too many features can be a resource drain and can negatively impact user experience. As someone in charge of managing a product and making sure it beats the competition, you spend the bulk of your time focused on developing and improving new features, but removing older features is just as important because all of your live features take up resources. By removing features your users no longer find useful, you free up valuable resources and can devote more time to improving your product’s user experience.
When usage is consistently low for 90 days or more, it’s time to look into the reasons why, and maybe consider retiring or improving a feature. Look into the maintenance costs of the feature, its relevance, the importance of the problem it solves, and gather customer feedback to decide if you need to roll it back.
Keeping a feature up and running costs resources, i.e., the money you spend on the infrastructure and the team that looks after the feature. If the maintenance costs are high and feature usage is low, you’re losing money just by continuing support.
For example, hosting your feature on the cloud using Microsoft Azure can cost you anywhere between $8 per month to over $600 per month per server. If your feature requires multiple servers as well as a few hours per week from your development and support teams, you’re spending thousands of dollars on it every month. You have to figure out whether the feature is paying you back through user engagement or not. If it’s a resource drain, it needs to be removed.
A prime example: Microsoft will retire Internet Explorer on June 15, 2022. Internet Explorer lost the browser wars a long time ago, yet Microsoft persisted with it for years. Now they’ve decided to focus on Microsoft Edge — their faster, more secure browser. To figure out whether you should persist with a low-usage, high-maintenance feature, survey two segments of your user base:
Make sure both sets of groups have similar engagement/usage rates, so you don’t end up comparing highly-engaged users with low-usage customers. Through surveys, find out the reasons why Group B doesn’t use the feature and why group A does. For example, you could ask questions to determine if Group A still uses the feature because it solves a problem. For Group B, you'll want to uncover if the problem is not important enough for them to warrant using the feature or if they've simply found a workaround.
If data suggests the problem your feature solves is not that important to your users, give your customers a heads-up before removing the feature. If you find the solution is necessary, you should take steps to improve that feature instead of removing it. In either case, the data you collect from your user base will tell you whether it’s time to retire the feature.
Sometimes you roll out a new feature that makes an older feature redundant. Once a majority of your user base has switched to the new feature and its adoption and usage is where you want it to be, retire the old feature.
In YouTube’s case, they realized the personalized recommendations on their homepage included videos users had already watched. The ‘new to you’ feature pops up after a user has scrolled down many times without clicking on any videos. It shows videos from content creators new to them. If this works well and most users enjoy this new feature, it’s only natural that YouTube will replace their homepage suggestions in part or entirely with this feature. It can also be a potential competitor for the ‘explore’ feature on the website.
Netflix is also trying something similar with ‘Play Something.’ The new feature uses Netflix’s algorithm to find and play a show or a movie that’s of interest to you based on what you’ve watched before. If this feature gains popularity, it’ll result in a replaced or modified algorithm for the ‘Because you Watched’ carousel. The ‘Because you Watched’ carousel also has suggestions based on your viewing history, but usually you scroll past it or keep exploring it until you find something or move to a different carousel. The ‘play something’ feature takes it to the next level by playing a show instead, so the show cannot be anything random — it has to be closely aligned with your interests.
In both cases, the new features are intentionally competing with older features, and their success will result in the older features being retired or merged into the new ones. This gives you complete control of when to remove a feature from your product.
While the decision-making process and the cut-off point will vary for each company, if more than 90% of your users have switched to the new feature, it’s probably time to retire the old feature. Best practices suggest publishing the announcement on your blog and placing a feedback form on the page so your users have a place to express their concerns. You can also collect in-app feedback and send out the announcement via email to make sure it reaches everyone. By providing your users with a feedback form specific to the feature being removed, you’ll be able to answer their questions and keep an eye on your customers’ reactions.
In some cases, a feature may not align with the strategic direction of the product. In that case, you’ll want to take it down so you can focus more on current and new features that’ll add value to the product.
For example, LinkedIn retired its Stories feature because it didn’t quite fit with the rest of the brand. At its core, LinkedIn is a social network, so the instinct to try the latest features becoming popular on other social media platforms is understandable. However, as they found with ‘stories,’ people do not use LinkedIn for the same reasons they use Instagram or Snapchat. They removed the feature while still working on a replacement feature.
In LinkedIn’s case, they knew they could remove the stories feature without giving their users an alternative because it wasn’t massively popular with their users in the first place. However, keep in mind, removing a feature without providing an alternative runs the risk of decreased net promoter scores and negative feedback from customers.
Since this decision isn’t driven by usage data, removing a feature due to a change in strategic direction should be handled carefully. This means announcing the removal a few months in advance instead of a few weeks, explaining your decision to your customers in detail, and making sure all your customer-facing teams (sales, marketing, and customer service) have the information they need to address user concerns.
If a problem becomes obsolete, the feature that solves it must go. A problem can become obsolete as technology improves, new solutions hit the market, or your customers figure out better ways to solve the problem. As your product evolves with time, it too will outgrow some features.
Industry trends will often give you a good idea of which features have become or are about to be obsolete. Keep an eye on market trends, listen to your customers, and adjust accordingly.
When you’re ready to roll back a feature, UserVoice helps you collect user feedback and segment it by customer satisfaction and net promoter score (NPS). This will help you keep track of customers who have concerns over a soon-to-be-removed or recently removed feature. You can also collect feedback on new ideas, recently released features, the strategic direction of your product, and more. To learn more about feedback collection with UserVoice, sign up for a demo today.