Eyes that do not cry, do not see — proverb
Your team is in a hurry to meet a deadline — they deliver the feature you asked for. It looks great in the demo. Do you ship it?
Sometimes you have to say no.
In a past article I talked about saying no to a product release, a product, a company, and potentially an entire industry. This is often the only proper choice when the product is harmful to its users, either long or short term. Not too long ago, I rejected an inquiry from a recruiter whose company helps the oil industry figure out where to drill to get around environmental regulations. This article is about something different.
One of my duties early in my career was to “bless” a release as ready to ship (at Ingenuity Systems). My responsibility to my team, our customers, and our shareholders was to release a product that delivered on the promised value. And we did.
Famously, someone at Facebook said you have to “move fast and break things.” Turns out that’s not always a good idea.
When you work in serious industries, for example as I do now at Castlight, your customers and users expect a quality product. As a product manager, it is my duty to ship products that meet customer and user expectations. Sure, they should “delight the user”. Even more importantly, they should not disappoint or mislead the user.
Product managers often do not understand how they can ensure that their products have proper quality. Isn’t it the job of engineering? Good question, and the answer to that is no. The entire team needs to build and ship quality. And the product manager takes responsibility for the shipment.
Product managers often forget to specify that the product must actually work. This is often assumed, but it shouldn’t be taken for granted. At demo time you are shown one or two examples. When you are working with complex features like those powered by machine learning or AI, how can you know whether it really works as it will be used in the wild? Here are three ways:
- Know the test plan, understand the test coverage, and review the results of testing in detail. If you have a dedicated professional tester (which you should if the product isn’t free), that person is your best friend.
- Test the product yourself. A lot. Every day. Try to break it.
- Sit with real users and refrain from helping them as they struggle to use your feature. If they can’t use it, if they do something “weird” and get stuck, that product isn’t ready to ship.
Asking for a quality product or feature, and refusing to accept something less, might look like moving slowly. It might delay your release. It may very well be unpopular. Saying no often is. But maintain hope and determination — as a product manager you want to ship a good product. Accepting less won’t get you there.