October 2009

In the the world of Scrum & agile there is an ongoing discussion about contracts. Main problem is clash between agile approach of flexibility based on adaptation and the world of RFPs, RFIs and fixed contracts. I’ve seen many good talks and presentations about this, including also on the last Scrum Gathering. One thing I don’t agree with is the way the problem of risk is being handled almost in all of them.

People usually start their talks with the notion that in the basic time&materials contract (the only truly agile contract if you ask me) all the project risk is on the client’s side and other contract types somehow move some of the risk to the supplier. Some go as far as to say that in a fixed bid contract (fixed time, scope & cost) all the risk is on the supplier side.

First time I heard this I thought this is clever, but over time it dawned on me that this “risk sharing” is as much illusory as is the security of the fixed bid contract. Here is why.

The single biggest risk anyone faces when they build something is that this something won’t fit the purpose it was intended for. Causes for this can be numerous: the needs might have been not understood properly or the idea doesn’t fit the market, the needs have changed or the thing built can’t be used because of defects. Finally, the thing being built may be delivered too late for it to be used as intended.

No matter how hard you try the biggest share of this risk is always on the client’s side, because it is the client who won’t get expected benefits in the end. You may add as many penalties and harshly sounding clauses to a very fixed and rigid contract as you want, you still won’t get away from this risk.

Imagine you’ve spent 3 months writing the initial spec, then next two months on RFP/bidding process, then contract negotiation, then two years on developing your complex software system only to discover at the end of all this that the product is not exactly what you intended, is full of bugs and is irrelevant because in the meantime the world moved on (the usual, flawed, waterfall process). What then? Yeah, you can sue the unhappy supplier all the way, but how much good will it do to your business? You can not pay them and even bankrupt them if they were not careful – will that repay all the lost opportunity? Make up for lost time? No way!

This risk is like a boomerang. I’ve heard someone define a boomerang as a piece of wood that no matter how hard you throw away from you will come back and smack you in the face. Very fitting. The harder you try to move away from this risk with complex contracts the more you will be hurt in the end. You don’t need rigid relationship based on distrust to tame this risk, you need an adaptive relationship with people you trust but can check all the time.

That’s exactly what agile offers: tight control of the product being developed in short inspect&adapt cycles. Instead of trying to write clever penalty clauses and waste time on lawyers clients can closely monitor the progress of the team(s) that build their software. They can spot any problems early on and correct them. Or readjust the backlog to changing situation keeping their product relevant. Or change the team.

I’d say there is way way less risk here. What clients pay with for this is their time and involvement – they have to see test builds at least every sprint, they have to speak to the team, readjust the backlog and overall stay on top of their project. But all the time they have all the tools and controls to navigate the project to a successful end. A fixed contract just kills all the outset.

So, next time when people will again define contract negotiation as risk pushing I’ll definitely interrupt them with my little boomerang analogy, because I feel we must all realize this is yet another (paper) illusion.

To wrap up my coverage of the last Scrum Gathering in Munich a couple of words about the last day and then some general comments.

After the lame keynote session on Tuesday I decided to sleep in on Wednesday and I missed Harvey Wheaton‘s talk. It didn’t appear particularly interesting on the mini-agenda we were given, looked like another boring life story. My friend Tomek went in, and thanks to him I know I’ve missed a very interesting talk, a real world experience of implementing agile & Scrum in a startup company run by Harvey. Well, next time I’ll be in for Harvey’s talk.

Another session I attended was supposed to be about agile contracts by Peter Stevens. To the surprise of both speakers this talk was squeezed into the same room and time slot with Regina Mullen‘s talk. In effect both had time to speak, but not to take many questions from the audience (crowded, as usually, on chairs and in between). Peter’s talk was a mild disappointment for me. I had a chance to talk to him a few times and I read his blog, so I know he knows a lot more about agile contracting. Maybe he just decided to cut his talk to leave some time for Regina. In any case, I didn’t learn much stuff that I would not know before, but I think it must have been a good introduction for the first-time attendees. One thing I’ll remember from this talk is his story about Zurich trams contract. And it also reinforced my resolution to finally write about my take on risk in client-vendor relationships.

Regina’s talk was, on the other hand, an unexpected, entertaining diversion for me. Though she did include her life story into it (like way too many speakers at the Gathering), she did it in a gracious and humorous way and her thoughts on making lawyers use agile were quite interesting, even though I’m clearly not in her target group (IANAL). Nice talk, I just hope she has a chance to present this to lawyers and indeed change their ways. Part of the challenge throughout my career in the IT world, especially as a manager in the last couple of years, has been finding lawyers that would really understand what we are trying to do before helping us draft contracts, user agreement, licenses and so forth. We need more people like Regina there – I definitely would love to talk to a lawyer who does know outright what a sprint is.

After lunch the best regular talk to attend was probably one led by Nigel Baker (I suppose so, as I know Nigel – it’s not possible to get bored when he talks), but tired of all the Scrum politics over lunch I decided to go to the haiku workshop led by Liz Keogh. I was amazed by the results. I really did learn something new I didn’t know before, and it was really easy. It took Liz a couple of minutes to get us to write haikus, and since then they keep on popping up in my mind on their own. Two examples of the ones I did and really like:

Worn green carpet
carries us all patiently.

Engulfed in our own little worlds
between iPod earbuds.
Autumn leaves.

(Haikus composed during the workshop and sent in afterward can be read here).

I resolved to try and write at least one per day. I think they are more than a diversion – they are a great tool to awaken our pattern-matching R-mode processing (read Andy Hunt’s book for more on L&R-modes). Haikus being by definition based on surprising matches can increase creativity by encouraging the R-mode “search engine” to produce more unexpected matches, useful not only for more haikus.

The event ending session was a big surprise. I expected some energizing talk, so I took out my camera to record it. Instead what happened was that after some small talk and a round of due thanks to different people by Tom Mellor a “wave” exercise was performed. It was so jawdroppingly stupid I thanked God I had my camera out so I had a good excuse not to participate in this idiocy without attracting attention. I kept on recording though and I wonder now whether I should put up this video or not, as I’m afraid it could only damage the reputation of the Scrum Alliance.

Interestingly, most people I talked with afterwards agreed it was stupid, but during the round after the “wave” they said BS like “powerful” or “energizing” – only me and one other guy were honest enough to say it was “silly”. To me it was a clear example that whoever was supposed to lead this session just didn’t have anything meaningful to say, so came up with this thing instead.

I think this failure illustrates a deeper problem Scrum & Scrum Alliance has. Scrum is part of the IT industry’s response to the failure of traditional, waterfall based project management in software development that that is called “Agile”. Scrum is a simple framework for managing requirements, work and team during software development projects. Yes, it doesn’t say anything about “software” as such, but this is where its roots are. In other words – it is a way to manage projects (because, I’d argue, Scrum projects are still managed, even if there is no Project Manager as such) created in our industry.

Now we have to decide – is Scrum just that or is it something more? Should we focus on merging Scrum with technical practices in teams and keep focus on software (which is what Ken Schwaber seems to be doing with his new training program aimed at developers)? Or should it be treated as a industry-agnostic project management methodology that could take on PMBOK, Prince-2 and other heavy methodologies everywhere? Or maybe should it be part of the soft-skills portfolio for coaches, HR etc. together with “waves”, collaborative drawing etc.?

The problem here is that traditional management methods work quite well in heavy industries they evolved in, and there are lots of very good people in the soft-skills department – usually better than ex-software developers discovering their “emotional intelligence”.

In any case a choice must be made before Scrum dissolves into mist. Ken, Jeff and Mike did agile movement a great service by laying out Scrum as a clearly defined, tangible thing as opposed to agile’s vagueness. Scrum was something anyone could start using and derive benefit from following its simple rules – because even ScrumButt was giving more productivity than waterfall. That’s why it was so successful – exactly because of its sharply defined edges and simple clarity. Now this clarity can be lost amidst empty motivational talks and silly exercises lifted out of kindergarten.

I know what I wrote will be harsh for some, but well, that’s what I think. And being a nobody in this movement, a nobody living in a place most of you never heard of I can afford the luxury of being honest. And of course one thing will stand: Scrum itself and other agile methods and practices. No matter which Alliance trademarks what you can use them and be better off in your projects.

To close on a more optimistic note some appreciations:

  • Thanks to Boris Gloger and his team: they not only provided real, freshly pressed juice and real, freshly brewed coffee but also filled in where organizers couldn’t in a truly agile fashion. Their idea of just copying the materials on pen drives and handing them to people at the end was brilliant. Plus he paid for the beer on Monday.
  • Tobias Mayer: another true agilst, when everyone complained about the lack of Open Spaces and the board said they will make them next time Tobias singlehandedly organized “guerrilla Open Spaces” the very same day greatly enhancing the Gathering experience for many. Also, it is Tobias who told me to go to the haiku session instead of political discussions – thank you!
  • Howard Sublett and Cory Foy: Scrum Alliance needs more of you guys, not just 5 hours / month.
  • Stefan, Robert and other German friends who helped me find my way around Munich and explained some customs during the beer evening.
    Everyone I talked with during the breaks.

Overall, I liked the event as a whole, I hope next one will be way better – and that clouds over Scrum Alliance’s future will be gone by the time we meet again “somewhere in Europe”.

(There will be one more post in the Scrum Gathering series – about Scrum Alliance as such).

Here are my short notes from Scrum Gathering. Overall this edition is slightly disappointing as compared to the last year’s.

First the talks quality is a bit lower than in Stockholm. Or maybe I did end up in wrong ones because I didn’t choose well. Nothing besides the talks titles & speaker names was announced so I could only guess.

Initial remarks by Jeff Sutherland were delivered with his usual zest, but seemed to be just an incremental update from his presentation in Stockholm and his other talks. Must have been quite interesting for first time attendees though. One thing I picked up there is that I really need to read the original Takeuchi & Nonaka paper that started all of this. And I think it is worth exploring what those guys are up to now. If their one paper started this whole movement, then maybe their other work is also valuable.

On to other first day talks: “Coaching Scrum Teams with a User Centered Approach” with Mike Sutton turned out to be much less interesting than I thought. Two things I picked up there were Prezi presentation software and a book about brain functioning – that is before I left 30 minutes into the talk. Simon Bennett did a much better job with his talk on applying game theory to agile contracting. Though his conclusions were certainly nothing new for me his exercises were interesting and kept the group involved. Also Roman Pichler’s talk was interesting, though his delivery style – slow, calm and quiet – is not what I like. He discussed anti-patterns for Product Owners – or common mistakes made by Product Owners. One point we didn’t agree on is what he called “bungee product owner” – I think I’ll write more on this separately.

As far as the first day goes, though, it was Mike Cohn‘s talk that was the highlight of that day for me. I recorded the whole talk and with Mike’s permission I’ll post it on-line when I get back.

Second day opened up with some guy from Sweden – Petric Palm – telling us his life story and showing quotations from famous people and nice pictures. Certainly not a keynote level talk. I used that time to do my e-mail.

Next, organizers switched sessions – the one I wanted to go to was replaced with a product owner panel discussion I wasn’t all that interested in. When I realized that I wanted to go to Erez Katzav’s session but the room was already so full of people it wasn’t possible to get in. More e-mail then plus an interesting chat with another guy who did not get inside for the Katzav’s talk.

After lunch finally the highlight of day two came: Serge Beaumont‘s talk on tools for product owners (tools meaning practices and mental tools not software tools or anything like it). Learning from the previous sessions I came 20 minutes before the talk so I didn’t have to sit on the floor. Serge talked about ways in which Product Owners can organize their work and collaboration with the team, especially around that part of the backlog that is not yet inside the sprint. Serge’s observation is that once team is running at high speed with Scrum POs quickly run out of backlog or they are not fully prepared at the outset. The cure is to define a READY state for items that can go into sprints – an equivalent of the DONE state at the end of sprint – then build a Kanban-based flow that feeds the backlog of READY items so that it is never empty. Great talk and I’m looking forward to seeing the slides posted.

Finally, the day ended with the Scrum Alliance board – or part of it – answering questions. I have to say that it was a depressing experience. Both Jeff Sutherland and Mike Cohn were not present, I think they have left earlier, which was a surprise. The four board members present were looking sad, devoid of energy and vigor as they sat collapsed in their chairs. Sadly, the Alliance’s support staff – Howard Sublett, Cory Foy and others – looked much more energetic and alert than the board.

Plus the style of board’s answers was like on a corporate meeting: very diplomatic, rather avoiding or deflecting questions than giving any straight answers. For example someone asked if the board will be elected by the membership rather than elect itself. The short answer was NO, but instead of saying so one of the board members went into a lengthy discussion which included a story about some football coach from his college or university or whatever.

In summary it is quite clear that the Scrum Alliance is suffering from leadership deficit. “Improvement communities” are a great idea, but they can’t be an excuse for lack of leadership. After all self-organization occurs around goals, and it is not clear what Alliance’s goals are now. This apparent lack of vision is normal after what happened recently, but it can’t be allowed to last too long. Right now everyone is, I think, willing to cut Tome & others on the board some slack and give them time to find the direction but this won’t last long.

Personally I think the Scrum Alliance should open up, especially to members outside of the “old boys club” (as someone called it), to move ahead. Pushing Scrum outside of software development is IMHO unrealistic and in places wrong. The Alliance should rather focus on improving the quality of projects inside our industry. There is still a lot to be done in this sphere. The complexity and importance of the Product Owner role, for example, has been only discovered over time. The problem of awareness in the industry that is reflected in everyone still wanting fixed bid contracts is an ongoing problem. I think everyone could add to this list.

Making the CSM exam a real test of knowledge as soon as possible would be the first step in the right direction. I think exams should be like PMI’s PMP exams – there are no certified PMP trainers, anyone can train or learn PMBOK on their own, but the exam itself is known to be hard and is administered in a controlled environment to ensure one can’t cheat with a book or search engine. Such a structure promotes honesty and certificate value plus it is much more open and fair. I think this is a good example to follow. In the long term a good structure of exams like this could even make CSTs obsolete – which would help the promotion of Scrum greatly. Right now the existence of this closed club with unclear entry rules is choking the growth, especially in the parts of the world where English is not the primary language.

To wrap up my rant a short list of organizational failures:

  • venue: rooms too small, at times crowded – people had to stand in some workshops,
  • lack of proper events folder or plan – short plan in the conference ID is good, but without a brochure with descriptions of talks and speakers profiles one is left guessing as to what talks are about, it is way harder to make good decisions as to where to go, amazingly this information was even missing from the Scrum Alliance’s web site even though all speakers did submit it,

  • lack of Open Spaces – that was a great idea, one of the greatest advantages of Scrum Gathering in Stockholm, don’t understand why it got dropped,
  • lack of free Internet – at 1200€ one could expect that (was fixed on the second day).

To ballance it with some organizationa positives:

  • all most talks in English,
  • food – good and abundant!

So Obama did get the Nobel Peace Prize… Amazing! They could just as well give him the Nobel Prize in physics – after all, he didn’t do anything there too but just as well may make a groundbreaking discovery in the future.

I think Nobel Peace Prize’s credibility is now completely gone as the committee could sink no lower. Giving it to someone just based on hope he will do something, just based on his slick talk even before he could really take any significant action, is so wrong and stupid it pains.

But it may be a good thing overall. I just hope the absurdity of this idiocy will help more people wake up from their media-induced slumber.