How we sunset features
In light of our recent sunsetting of credit card purchases* (via Simplex), we’ve written a guide on the processes involved in removing certain features from products.
* Note that the credit card feature was only available in certain countries; we still support credit card deposits in Nigeria (via PayU). We’ll keep monitoring data and our customers’ needs and could in future add credit card purchases to specific countries again.
So, you want to remove a feature from your product? Let’s zoom out a little, take a big breath and break this thing into smaller pieces.
Step one: review the data
We don’t like to do things based on gut decisions at Luno. Most statements, features, spreadsheets and meetings only happen after a significant amount of data has been collected.
If there are fifty data points we look at before adding a feature, we’ll collect a hundred before removing it.
Data on customers
This part is arguably the most important: when we decided to add something and it was used by some customers, it means that they found that feature (or all of our features) useful. We need to understand who these affected (and potentially affected) customers are.
These are some of the things we got answers for, in no specific order.
- How many customers will be affected?
- What percentage of overall customers who use(d) the feature?
- Did they try it once or did they use it all the time?
- What percentage of potential customers did not use the feature?
- Why did those customers not use it?
Data on alternatives
- What are alternatives available to customers?
- Do we provide a superior user experience or service to the alternatives?
Data on company metrics
- What resources (such as engineering, customer support hours) does the feature require?
- What is the lifetime value of customers who used it?
- What percentage of overall product use and revenue comes from the feature?
- What is the expected cost of keeping it vs. the potential losses from removing it?
Is the product adding or distracting from our goals?
Step two: developing a sunset plan
The type of feature (or product) that will be discontinued will determine the detail of the plan, but in short: it should be very detailed.
We use a simple, shared Google Spreadsheet, broken down into sections (such as Engineering, Community, Marketing), each with their tasks to be done, status and a timeline for completion.
Customers and staff are going to be affected and it is irresponsible (and expensive) not to plan properly.
Some things to consider:
- What are the timelines: how long will the feature be available, how much time do the Engineering team need to remove it, when will we no longer be able to provide support?
- Who owns the process and writes the step-by-step document?
- What are some of the post-sunset questions that may arise and is everyone briefed?
Step three: communicate
Since we know who are going to be affected, we can create a more targeted message and deliver it through the most efficient channels.
If you are removing a feature that is only going to affect customers in a certain country, don’t blast your entire mailing list with the news. To everyone else, it’s simply irrelevant. You can often segment it more and just communicate to affected customers (those who used the feature in the past) and potentially affected customers (those who, based on data, seem to be likely candidates to use the feature sometime soon).
First, we involve all of the Community Team, since they are the first line contact with our customers. They might be fielding some disgruntled customer emails and phone calls, so the more information they have, the better equipped they are to empathetically respond.
Depending on the type of feature you’re removing, you might want to create easy-to-read canned responses your agents can use to respond to customer questions.
If you have support documentation or a Help Centre, be sure to update your articles to indicate that the feature is no longer supported.
Reach out to customers with a simple, concise message.
Some things that might be useful:
- Give enough detail (or link to further resources) so that nothing is unclear
- Give clear dates and times the feature will be discontinued
- Recommendations on alternatives (if relevant)
- Have empathy: you’re removing something that they might have found useful
We notify our customers through different channels: emails, push notification, in-app message, phone calls. Again: with proper data and segmenting, you’ll only need to reach out to those who are affected and can reach them through the most effective and least intrusive channel.
Feedback and monitoring
The days and weeks after sunsetting a product (or announcing that it will be discontinued) are crucial to monitor sentiment and make sure that the message received was the message you sent. Our Growth and Community teams gather feedback and data from multiple sources, some of these you can consider are:
- Internet forums
- Comments on social media channels
- Responses to surveys, phone calls and emails
It might seem easier to “just keep” a feature that isn’t used by your customers. The best product managers, however, are the ones who are great at saying “no”, those who are slow to build and quick to kill.
Often by adding something, you may be diluting your overall product, but inversely, by removing something, you remove clutter and distraction for your company and your customers.