Brave new world?

Leave a comment

Organisations should be interested in Scrum adoption, because there is something to gain. Incidently a reduction of pain, would be concomitant. So what are the common benefits?

The benefits are significantly better quality and productivity, but there  are a handful of other important benefits(possibly even more important):

1. Focus (reduced re-work/confusion, therefore happier floors)

2. Higher possibility of building the right product

3. Speeding up learning, which increases the rate by which previous benefit is accrued.

Coming back to the topics of productivity, I’ve personally helped teams increase it in the range of 30 to 60% and am convinced that those teams can do even better. So the claims of a 200% increase in productivity by some others are not necessarily codswallop. (However I’d caution against productivity becoming the MAIN/ONLY goal of adoption.)

It is surprising how few organisations are able to realise these benefits; this is often due to management not really holding teams (the WHOLE team) accountable, and, not creating an environment where quality is key and teams, SM and PO are empowered.

It’s a brave new world, if we can get there.

Why do we need self-organisation when we have micro-management!?!

2 Comments

Good question!

I’ll try to explain, while avoiding a close coupling with Scrum, though some tangential references may well slip through. Let’s start by considering something which is going to be familar to the gen-X managers: Pseudo-code, a key step in structured software development (or defined s/w dev processes). As the design phase is winding down, the last level of detailing was being filled in by way of pseudo-code. Pseudo-code: a detailed description of steps of a computer was supposed to execute, but not in a programming language. This was to be simply translated into code, by a worker bee (programmer) into a computer language, viola, ofcourse we have a flawless system. The need to write pseudo-code for someone else (the worker bee/mule) is, an admission if only latently, that we are hiring dimwits. If we follow this train of thought, another conclusion looms: usage of psuedo-code is a strain of intense micro-management, and a failure to understand the real nature of programming (at least programming in the small). As an aside: devotees of big-upfront design may have diagnosed correctly, that programming in the large, brings about its own set of problems; However they have unfortunately taken the wrong pill. They can detoxify, by reading Jack Reeves thoughts on software design. BTW, the gentleman, has nothing to do with the agile jamboree, just a very clear sighted (an endangered species) thoughtful, software developer. There is nothing wrong with a limited amount of upfront design, as long as we don’t try to develop the ‘perfect’ solution, while keeping in mind that this upfront design is just a draft, which can only be final, when the software works! (testing time, anyone!?!). So, unavoidably, we have to grapple with all sorts of detail, where the devil is hiding. So who is going to do all that grappling?

This brings us nicely to consider micro-management of teams. Serious software development takes place in a far more complex and fast changing environment than ever before. The work is highly interconnected with frequent changes (and surprises) streaming in. Many competencies are involved, with many things having to come together for a a successful result. It is impossible for one manager to do all the basic thinking and detailing. Much of the simple software has already been built, and most worthy teams are left with the implementation of involved software solutions. Therefore, the manager in question has to continually re-issue instructions to the team as events occur, surprises spring, lessons knock hard and the real target (software we need, as opposed to have wanted some time ago) reveals itself. It is, I’m afraid something of a losing battle.

Instead, the manager must work more as a facilitator, who ensures that the necessary resources and tools are provided to the team, and that impediments to team deliverables are removed. Having done this, it is best for the manger to get out of the way. It is not the manager’s job to order the team around; rather, it is the team that decides on how much to commit and how to deliver (from the top of the product backlog).

Breathing down the team’s neck and micro-managing it from the outside sends a signal that the team is not responsible. It then becomes the manager’s job to commit and then worry about how to get things done. This limits productivity, innovation and creativity in the team, chokes communications and, in time, results in disengagement and apathy. Actually this state of affairs is so widespread that it is the new normal.  That is why, we should encourage self-management.

If ownership firmly rests with the team, there is greater focus, sense of responsibility and motivation to perform. Let the team manage itself. The manager’s job is to keep the focus on the bigger picture and help if the need arises. At the same time—and paradoxical though it may seem— the manager must not lose sight of the critical details (important when teams are dealing with the rest of the organisation).

So, get your team together, emphasise goals, facilitate learning, offer to help, make it clear that you are watching and then,…let go! You should rather be spending your time to prepare for a role at an n+1 level than get bogged down with the details at n-1. Potentially a depressing corollary, is that one circumstance micro-management could succeed is where the project on hand is relatively straightforward! So, maybe, that project you are so successfully micro-managing, is just a run of the mill work, where low intelligence finds a comfortable home. It would also be interesting to know what you think of the BBCs advice on micro-management. Actually if you think about it, many managers cannot even really micro-manage, but try to give the impression that they are on top of things. A waste of energy, time and in acute cases, even space.

Therefore, in general, I advise eschewing micro-management, but hold your breadth, further flutters await you….

Some thoughtful birds in my circle of acquaintances, debated over the need for managers at all (in all shapes and forms !?!). I’m all set to write about that as well, in a day or two.

Reverse Viennese Syndrome

2 Comments

Happening to read this, I was struck by how many people (within ICT industry)  in India have the reverse problem. As a Viennese resident remarks- “If you live here all the time, you have nothing to compare it to – and you don’t know how good you’ve got it.” or to paint a picture for a reverse problem “…. and you don’t know how bad you’ve got it.”!

I relate to this due to a similar personal experience on my first visit to foreign shores (London) when a local remarked how polluted London was. Arriving from Bombay (Mumbai), I remember thinking the poor fish lacked sense of any sort. But maybe it was me who lacked perspective.

What has all this to do with Scrum? Many, many people have no idea of how a well chosen and tended team can feel, perform and delight all. So many projects are in a death march mode or in an apathetic semi-daze for so long, that such a state seems normal! Just as the Viennese are inclined to overlook how well off they really are in their city, the s/w dev teams here are not exposed to how well (comparatively) things can be run.

Is this a plug for Scrum? Partly, but mainly it is an attempt to show that there are much better means to live within a software development project.

Is this relevant to the times? Yes, in spite of  all the huge numbers of people claiming to use Scrum, the percentage who do it properly is very small indeed. Scrum-but will not make your life better, Scrum will.

I’ll attempt to provide a sense of how one can help people see a brighter horizon. Most software development teams are under pressure to deliver on very optimistic estimates of poorly thought out functionality. On top of this teams will quickly settle into a habit of creating poor quality functionality.  All this leads to continuous rushed activity that results in comparatively little progress.  This is why people spend late nights in office. How they use their time in office is another story.  However too much overtime over extended periods will be accompanied with lower mental efficiency and strong tendency to make mistakes. Why do you think in competitive sports each team tries to put the other under pressure? To encourage mistakes of course. What are managements doing when they put pressure most of the time?

First thing you need to realise is that a normal day needn’t be tense and rushed. It can be comfortable but purposeful. Good planning to deliver a reasonable amount of functionality will provide best results.

A cross functional team results in testing being done very close to development, which means that we get a lot of feedback within a sprint. Also multiple perspectives help in detection of mistakes earlier.

More importantly we get serious external (customer/customer-representative) feedback at end of every sprint. Therefore we come to know how we are doing from a very early stage. At every step of the way we have a chance to improve at a reasonable pace. This is almost a dream in a waterfall project. Everyone pretends they are already very competent or even perfect. Then at the end, when they all are floundering, and then it doesn’t end….

Where as in a a well tended team, the support every one gets  is very enabling,  a good ScrumMaster is there to see to it. It provides a high level of ‘safety’. This supports people to very quickly get better, be open to learning from their mistakes. These may or may not be mistakes of individuals, but a culture of blaming one another is strongly discouraged.

To summarize: A good project delivery structure (delivery and feedback every sprint), a supportive cross functional team and picking up a reasonable amount of work per sprint will together result in a steadily productive, yet comfortable feel for Scrum projects. Do it properly and you’ll then realise how badly you’ve been having things for so long.

An extended entertaining tete a tete

2 Comments

I had the pleasure of meeting Jeff McKenna last weekend. I took him around the city a little bit and had lunch with him as we chatted about various things, with discussions about Scrum taking up about half of the time. He was very engaging. He started programming in the sixties as a young teenager and talked about a seven pass compiler! Blimey! Glad I didn’t have to go through that wringer. Must have been fun in someways, but surely not pure joy when an error was thrown up on the sixth pass. He was also among the vanguard of the Smalltalk community. That is why meeting him was also a privilege. Such a spectrum of experience! He has I’m sure a store of stories to narrate and observations to make, only a small portion I could listen to.

In any case, one of the points he made really well was how you can always break a seemingly big story (feature) into much smaller parts. Many people who come from a traditional background seem to struggle with the idea of taking a functionality from description to implementation in just one sprint.  They think “well, our application is complex (glossing over the fact that complex and complicated are different) and we can’t implement anything meaningful in a two week sprint”. Indulging in some levity he makes the point thus “you can break it down to a few keystrokes!” Touche! But the point is serious. You can take it from people who have seen a really wide range of computer software applications, many times more that the average Mac, that one can always, always break functionality down to a couple of days of work. Surely within a development environment this is possible. There may be other organisational barriers that mean that the implementation cannot be taken into pre-production (or whatever) within a two week timebox, but the point of Scrum is to remove as many of these barriers as possible.

More on this later .

Strictly left footed right wing striker

2 Comments

Background: The Pune vodQA 2012, where I was involved, but not entirely committed(A frivolous reference to Scrum lore, about commitment, chicken and pigs) in more ways than one, was interesting instructive and illuminating! Read about it here.

Mr Chaitanya Nadkarny, a senior member of Thoughtworks consulting gave an opening talk, covering among other things which factors are going to affect the lives and careers of testers. One point he made was with regard to how testers can be generalizing specialists, or specialising generalists. It took some time for the audience (at least I did take time) to get it’s collective head around this.  Mr Nadkarny’s point here was that

One could be a performance tester (specialist), but also interested in expanding capabilities (therefore generalising) to do general functional testing, manual etc.

OR

A tester (generalist) now wants to be more skilled and focus on an area of testing, for example usability (specializing).

MAIN POST:

Since I see the world through Scrum tinted glasses, I’ve been struck by how Nadkarny has put in a nutshell the possibilities of how people develop, become more valuable and competent. Many years ago a programmer was doing most if not all activities in relation to developing systems/applications, including testing. As the field developed people became more and more specialized. Eventually in the noughties, we have extremes of specialisation. Even being a tester isn’t enough of specialisation, one is now a manual tester or an automation tester, or a performance tester. The side effect of the industry moving in this direction is that people are squeezed into ever narrower roles, even though this isn’t required and actually creates inefficiencies. Finally we have the equivalent of a striker who can only play on the right wing,  kick with his left foot, but will not head the ball into the net, even if the ball is at eye level, straight ahead and the keeper is out of position!

This attitude is so entrenched  and pervasive in some places that even managers have often fallen into the trap. People are type-cast and given no opportunity to learn new things. The Scrum team, essentially being cross-functional, would struggle to succeed if a group of super-specialists have to work together, even if they are initially willing to collaborate. Why so?

Firstly: There is little understanding of what the other person does and needs.

Secondly: Load balancing at different points in a sprint is a nightmare, as work of a given type isn’t evenly loaded throughout the duration of the sprint.

Thirdly:  Majority of work in any organisation is done by generalists, because much of the work doesn’t need a too much of specialization. How often have you visited a general practitioner vs a specialist?

Lastly and very importantly: it encourages the continuation of waterfall habits.

Time to generalize. A cross functional team would need generalizing specialists.

On the other hand, once they get there, they might then need to be specialising generalists, as well… and so it goes.

Developers are people too!

1 Comment

I happened to come across this absorbing article: Whales are people too! and in then the word speciesism (a one word tongue twister, say it out loud, quickly, twice in a row!)

Now, I’ve lost of count the number of managers/executives who proudly say that their organisations are doing Agile (whatever that means ) and refer to people as ‘resources’. In light of Mr D. Adams’ observation (that humans are the third most intelligent species on earth) it is possible that these enlightened managers are right after all. We are all resources.

However, on a serious note, it is worth considering, since dolphins have a strong claim to rights, if our developers also need some rights.  Also more pragmatically, always keeping  this in mind makes management effective, and true; Not a wished based glossing over the real people development and attitude issues and avoiding necessary but difficult conversations until it’s too late. A team of people, are not simply a summation of individuals, and their interactions can easily be unpredictable, someone leaving the team, could result in a productivity increase, or even velocity increase. If only developers were people.

N.B: Prima facie this is tangential, but on second glance quite central to Scrum. In this article I’m not being sarcastic, but tongue-in-cheek. (Note on N.B to some readers: These things have to be explained to most ‘resource managers’  in the ICT industry)

What can be done when the PO isn’t empowered?

Leave a comment

A fairly common situation in organizations is the presence of a PO without authority. It is common for a BA to play this role, and mainly due to the mistaken belief that functional knowledge is  the only criteria for a PO. Also the term ScrumMaster suggesting an all important person in a move to Scrum, contributes to this thinking. As in many things the best course of action is to preempt this situation by educating the rest of the organization (before) the start of Sprint about the roles of SM and PO. Once the realization set in, that the SM is not responsible for the delivery on a sprint to sprint basis, and in terms of delivering the product to the customers, people outside will turn to the PO. The PO will then slowly take on a stronger and more meaningful role, suffused with responsibility.  In my experience, coaching project teams and organizations, this take some time. The first level of understanding come about in 3 to 4 sprints, but deeper understanding typically takes a release or two (maybe 3 to 6  months). What one can do to make this more visible:

1. Re-iterate that Scrum is a radically different system of running software projects.

2. Encourage interaction between team and PO, as well as PO and external customers/client.

3. Explain to all concerned that while the team is largely responsible for “building the system right” every sprint, the PO is reponsible for making available “The right system”. While the team is collectively responsible every sprint, the release to the customers is the responsibility of the PO.

4. In addition for a release the PO is ultimately responsible for delivery of the “right system right”, in other words the appropriate system which is shippable (i.e of high quality)

Older Entries