The Agile Executive

Making Agile Work

Customer Driven Testing

with 4 comments

David Anderson, James Shore and I engaged in a Twitter dialog about the post Every Nine Minutes. Our exchange highlights the importance of balancing development, deployment and operations, as follows:

what concerns me about continuous deployment is it is developer centric. Deployment has a cost for the customer [DA]

Continuous deployment involves automating what must be done anyway. Unless you mean end customer, don’s see the costs you mean [JS]

yes end customer. not everyone wants the UI changing every 9 mins 😉 what about training? marketing? etc..? [DA]

I think that’s an issue that’s raised by, but independent of, continuous deployment, and quite manageable. [JS]

 I want 2 c people talk about this holistically. If deploying more often is better then how do u reduce cost/impact on customer? [DA]

It seems to work quite well for their clientele. Their iterative customer development work might be the ‘secret sauce’ [IG]

is this an aspect of their tech enthusiast, early adopter market? [DA]

I think it is deeper – they adjust their testing to the needs of their market segment. Will blog on it later today. [IG]

The software IMVU produces is similar in some respects to what Clay Shirky calls Situated Software: “… designed for use by a specific social group, rather than for a generic set of users”. . .  IMVU’s  testing seems to be in good accord with the needs and priorities of their young users. Certain deficits in their software do not seem to be terribly important to their clientele. As pointed out by Elizabeth Hendrickson it might not be perfect but the software does the job for the target clientele and creates value.

The way IMVU develops the customer in an iterative manner (in parallel with iterating on the product) seems to be the key. Deep understanding of customer and problem determines testing strategy. Given their Lean Startup orientation, “good enough” testing seems to be quite appropriate for their business design:

  • Product release cycle in hours, not years
  • Tightly coupled with customer development
  • Minimum feature set, maximum customer coverage
  • Rapid hypothesis testing around market, pricing, customers,…
  • Extremely low cost, low burn, tight focus

IMVU is an example of the approach advocated in Agile Considerations for CXOs: don’t harness Agile into a rigid business design; instead, develop a business design around the capabilities of Agile.

Written by israelgat

March 15, 2009 at 3:28 pm

4 Responses

Subscribe to comments with RSS.

  1. This has been my experience with the chaps over at 37signals (Basecamp, Backpack, Sunrise, etc). Focus like a laser on the 80% function, strip out all the ‘nice-to-have’s’ and, based on my observation I would add another dimension– iterate only when you are absolutely certain the functionality will improve the experience for the vast majority of the so-called social group. -Dave from wikibon.

    David Vellante

    March 16, 2009 at 5:44 am

  2. These concept are being defined right now.

    The approach IMVU has taken is only viable in a narrow segment of non-critical social applications. Worse case scenario is someone’s IM avatar isn’t right. Look what people will put up with on twitter.

    To me continuous deployment is a holistic approach to get the fixed price of deployment as low as possible. The lowest hanging fruit is fast and fully automated deployment with no noticeable user impact. This doesn’t mean that deploys have to happen as soon as a developer commits and the tests pass. The second thing to work on, and this is where the criticality of the system comes in, is to lower fixed price of qualifying the software as this is often now the most visible bottleneck in the process.

    One thing to note, from a technical perspective the impact of the act of deploying can easily be brought to nil. From there the impact on the user mentioned by DA can be lessened by designs that offer ‘least surprise’ being consistent, usable and well documented.

    Marketing is another beast. Of course, context matters and for some business models and approaches there might be a benefit to not releasing features for marketing impact. I’m not totally convinced this is necessary, but it is understandable strategy, even if it is rooted in the past.

    Andrew Clay Shafer

    March 18, 2009 at 12:09 am

  3. According to Caper Jones [2008], MIS software accounted for 50% of all US software in 1990; 43% in 1995; 29% in 2005. We are obviously witnessing the rise of new types of software such as web applications. (Jones actually believes web applications could account for 50% of all applications in ten years).

    The question on my mind is segmentation of applications according to the way they are used by specific clienteles. Coupling such segmentation with testing under the “good enough” criterion is intriguing. IMVU seems to be very adaptive in this respect. Can such adaptivity lead to new economics of
    software testing?

    One other noteworthy point: important that cost of deployment is, the bigger question is life cycle cost. IMVU’s readiness to roll back quickly might generate signinficant life cycle cost savings.

    Thoughts please!

    Israel Gat

    March 19, 2009 at 11:19 am

  4. […] manner. Consequently, the company’s business design is likely to be transformed. Click here, here, and here for more detailed discussions how the business design gets […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: