Tuesday, April 22, 2008

Your Business Case for Agile


My friends at Innovel have a blog entry titled "Build Your Business Case for Adopting Lean Agile". Take a look.

As you try to get a revolutionary idea adopted, remember that you must always be selling the idea (see John Adams to the right). (Note: You don't even need 50% of your countrymen to agree, if you would succeed.) Your idea may be brilliant, may even be one of the best ideas ever to appear on this planet, yet you must still sell it.

To start anything in business, you need a business case. And given that everyone has a say and an opinion, in fact you need to have a good feel for all the benefits involved, so you can explain the WIIFM to anyone. (WIIFM = What's In It For Me)

And given that change is happening all the time, you even need to keep your business case up-to-date, since Agile will always be challenged with some equivalent of "what have you done for me lately?" People change. People forget. Attention gets re-directed.

So, what is the case?

Well, coincidentally Israel Gat of BMC is giving a talk this week at Agile-Carolinas (and another at APLN-Richmond). "Leading the Disruption". The idea that for BMC, Agile was so successful that it was disruptive, particularly for downstream processes. Such as sales and marketing. In a prior presentation he had shown hard data that to me proved that Agile had given them double the productivity for a large IT shop, compared to essentially waterfall (compared also to what they were doing before).

There are many people to whom we must make the case. Let's assume for now, to a senior business manager (maybe the CEO?).

BMC so far has doubled productivity. Umm. Impressive. In fact, Scrum (the main flavor of Agile that BMC use) was built to increase productivity by 5x to 10x. And it has been proven to have done so with some teams (the community has data in various papers). We don't (yet) have the broad based data to show whether "some" really equals many, most, likely, X%, for certain types of teams, etc. , etc. Based on the anecdotes I hear, most teams are clearly better even when not doing "most" of Agile. And teams that adopt Agile well are typically very much more productive.

Let's look at this a different way. The multi-functional IT teams (including business people, etc) are producing: more, faster, cheaper, and better. All at the same time. As in, higher quality, more accurately what the business needs (not just what they said), faster time to market, etc. What's not to like?

Another benefit senior managers like is visibility. They can tell whether an effort is on target or having problems. (More on this later.)

As Jeff Sutherland says, the ability to change directions quickly, and to focus this fire hose of productivity on a competitor's new feature is very valuable.

As an aside: One thing not to like is people doing cowboy agile. That is, people saying they are doing agile, but who really are not. At the extreme, this means people whose definition of agile consists of "we don't do documentation". Cowboy agile makes your audience more skeptical of real agile.

OK. One more thing on this topic.

How do you prove this for yourself?

Take people to the Gemba; show them the team room, the team, the product owners of successful teams, the working software. (Gemba is a Japanesee word, famous in Lean, meaning "the place where the truth may be found.")

Also, gather metrics from day one. I will talk about other metrics later. Today let's talk about velocity. Gather velocity (the sum of the story points of Product Backlog items completed in an iteration). Show that it is reasonably consistent from sprint to sprint. Then show the velocity increasing. Remove some more impediments. Show it increasing more. Impressive. A blind man can see this stuff.

If the velocity increases by 50% and the business value of new stories is not diminishing, then clearly you've got something going on. All you have to prove is that Agile is a good investment compared to other investments. Not hard if you work at it.

Will every team succeed? No. (Actually even failure is a success with Agile...topic for another post.)

Can fairly good people generally show impressive progress? Yes!

No comments: