If the Customer is King, the Product Manager is Regent

On Not Bending to Customers' Whims


by Katharine Nevins (blog: http://katharinenevins.posterous.com/)


The goal for any startup (or product manager for that matter) is to build a product which customers need and love.  It’s easy to call yourself or your company customer-centric, but actually doing what is best for customers isn’t always straightforward or intuitive.  There are two main reasons why listening to customers doesn’t inevitably lead to good products.  Firstly, not all customers need or want the same thing.  Secondly, customers often do a poor job of knowing or articulating what they want.  Fortunately, both of these problems are addressable.

For most products, different customers will have different needs.   Power users have different needs from new users: Sarah Dillard points out that Quora’s integration of feature requests from its power users have led to a bewildering experience for new users.  Different user segments may also have conflicting needs.  Natasha Prasad gives a great example of the trouble Digg had appealing to its original “geek” segment while trying to attract more mainstream users.  While working on Quickbooks, I saw a constant tension between adding features for power users or new verticals  vs. upholding our reputation for ease-of-use.   Creating new SKUs for power users (e.g. Premier edition) and verticals (e.g. non-profit edition) helped, but maintaining too many separate versions of a product is not ideal for a lean startup.

The most elegant “lean startup” solution I’ve seen to this problem is to use partnerships and open APIs to extend functionality.  Eventbrite, for example, allows third-party developers to build features or niche solutions such as an event check-in app through its apps showcase.  Eventbrite remains lean by allocating its resources to only the highest-impact new features.  Its customers all benefit by getting a simple but powerful product with an option to add only the extras they really need.

This API-based  approach has famously worked for products such as Salesforce (with the AppExchange) and Apple’s iPad/Pod/Phone (with the App Store).  Social products relying on user-generated content can make this approach work as well.  Apps built by Twitter’s developer community allow different customer types to use Twitter as a marketing platform, a news feed, or a social tool as they prefer.  The (potential) challenge for newer companies such as Foursquare and Quora is that they’ve already built their products for an early-adopting niche.  Now, they need to remove or change the early adopters’ favorite features so the product can be used by Normals.  Is there an elegant way to roll out a “Krunk-badge”-free version of Foursquare to soccer moms while keeping the existing, bar-hopping user base happy?  I’m not sure.   Thoughts?

Misleading or nonexistent customer input is even more unintuitive to deal with than conflicting customer feedback is.  Even within a specific customer segment, customers often can’t or won’t tell you what they really need.  Customers often ask for features they wouldn’t actually use.  Watching focus groups talk about their personal finances from my side of the one-way glass at Intuit, it seemed to me that their well-intentioned requests for better personal finance tools would in no way change their actual spending behavior, and therefore the tools would probably not be used even if we built them well.  Other users don’t ask for features they would use.  Prof. Piskorski’s research suggests that one of Facebook’s biggest uses is by men looking at pictures of women they don’t know.  However, these men probably wouldn’t have asked Facebook’s product managers for a better way to check out girls, and it’s possible they wouldn’t admit to this behavior if asked by a researcher or product manager.  Finally, customers are limited in the solutions they think to ask for.  A customer who didn’t know about email wouldn’t ask for a BlackBerry.

Separating
what the customer needs/ wants to do (the job) from how the customer will do it (the product) is the key to addressing this issue.  The customer often knows what job he wants to do.  If he can’t articulate it well, then observing his behavior as IDEO does will help the product manager figure it out.  The goal of the product manager or startup founder is not to tell the customer what to do, but to identify how to help the customer do something he wants to do anyway.  Customer statements of how they want to achieve their goals should be taken with a grain of salt.  In the year 2001, I would have asked for a faster Napster when what I really wanted was Pandora (though I didn’t know it yet).  To identify the best how (product), the product manager should come up with a few alternatives to test on users, then use their input to decide.  In this way, she can create a product even better than what users would have asked for themselves.