Saturday, June 20, 2015

Short Scrum and Long Scrum

There are many many ways to play Scrum. In some sense Scrum is infinitely flexible.

But let us pretend for a moment that there are only two types of Scrum, which I will call Short Scrum and Long Scrum.

Short Scrum

In Short Scrum, maybe where a Sprint is completed in a day, we do not have a long Product Backlog. Maybe the Product Backlog only contains enough work for one day or maybe two days (one or two Sprints). Imagine also that a ‘whole product’ (however we define ‘product’) or release can be completed in one or two days (one or two Sprints) or less.

With the Sprint length of only 1 day, we might have ‘Daily Scrums’ every hour (e.g., over an 8 hour day). The Sprint Planning Meeting might last 20 minutes. And the Sprint Review at the end of the day might last 10 minutes. And the Retrospective again might be 10 minutes.

Short Scrum would be useful when a Team has lots of small things to do, and those things are changing quickly (e.g., daily, and maybe being identified during each day). (For now let us assume still that the Team is the ‘ideal’ size of 7.) And you want feedback very quickly (each day). And the Team can build ‘working product’ (something that can be demo-ed that is in some sense complete) in 1 day most of the time.

Obviously ‘Short Scrum’ is a special situation or set of situations. Still, there might be many situations in the US (or the world) like this. Thousands or millions.

And, obviously, some situations might be similar to this, but not have exactly these numbers.

Getting the feedback every day can be enormously helpful.

One classic example is a small family doing a set of chores over the weekend.

This model also implies a high level of interaction in the team, during the day.

Also note that the concept of ‘long term planning’ has little meaning. Maybe long-term planning might be planning for the week. And note that Scrum (the simple framework) does not mandate any Release Planning.

The Velocity concept might be minor or not there, especially if it is not useful. For example, if the

Team is taking on a different type of work each Sprint (each day), then Velocity is unlikely to be consistent. Also, since a ‘release’ is happening every day, there is probably less need or no need to have the forecasting knowledge that velocity can give the Team. Still, in terms of continuous improvement and perhaps for other reasons, velocity might still be useful.

Few or none of these ‘criteria’ are ‘hard and fast’, but one certainly can imagine situations like this.

Long Scrum

In Long Scrum, imagine the Sprint is 2 weeks long. And imagine that it will take several Sprints (e.g., 8) to complete the product (meaning possibly: release the first release of multiple releases).

The Product Backlog might be very long (e.g., 6 months of work or more), and might include items that we might never get to.

The Sprint Planning Meeting might be as long as 4 hours (although usually somewhat less), and a Sprint Review might last as long as 2 hours (if useful). And the Retrospective might last 90-120 minutes, and of course help the Team increase their velocity.

And we have the ‘normal’ 15 minute Daily Scrums.

This type of Scrum is more useful when the Team has a big mountain to climb, and has some understanding of the mountain beforehand. (The mountain is a ‘product’ with a decent vision of what the product will become. The product is anything that that Team might ‘build’… tangible or intangible, physical or electronic or ?. But we do want the pieces to be demonstrated.)

Imagine that the Team has the worked divided into 8 small pieces (small stories) for each Sprint. And it takes some time to get each piece to be working. But, at the end of 2 weeks, all 8 pieces can be working. And it is useful to get feedback from business stakeholders: ‘do you want what we have built so far?’ And the right people will come every 2 weeks (in part because we have built enough to interest them).

The Product Backlog might still be changing, but the basic scope of the product is not usually changing a lot, or at least we do not expect it to change that much. We at least think we know what to build each Sprint, although in the Demo we sometimes discover that the business stakeholders do not like some pieces (some stories).

In this situation the concept of long term planning (well, planning over the 6 months, which would definitely seem ‘long term’ to the Short Scrum team)… is useful. And needed, necessary. In part because the business and the customers and the marketplace require some relatively good prediction of when the product will arrive or that ‘it’ will arrive by X date. Time is quite key in this way, at least usually.

So, Velocity is more needed. And more do-able, in part because the work is relatively more
consistent. And the Team needs to stay together, so they have time to discover their real velocity.


I have three main points in describing Short Scrum versus Long Scrum.
1. Scrum can be played many very different ways, and yet all can be called Scrum.
2. Some of you will mainly be playing Short Scrum….where the work is changing frequently and being identified only a few days before it is done. Still, you can use Scrum.
3. Short Scrum can seem very different than Long Scrum (to both sets of people), and yet both are Scrum. There are also, of course, varieties of Scrum that are some hybrid of short and long.
Hope this is helpful.
Your comments please…

No comments: