Dialogue Sheets Revisited
In January last year InfoQ ran my piece on Retrospective Dialogue Sheets. By the end of the month nearly 1,000 Dialogue Sheet downloads had been served by the website. A few months later I had a note from one of the editors saying it was one of the most read pieces on InfoQ.
The sheets have proved themselves popular at conferences and community events and I regularly receive e-mail from people who have been introduced to the sheets, use them and find them useful. A year on it seems appropriate to take stock and report on how the sheets are being used.
What is a dialogue sheet?
Dialogue sheets are design to promote discussion around a particular topic, in this case retrospectives. For those who haven't read the original article, a Dialogue Sheet is a large A1 sheet of paper around which team members seat themselves. The sheet is pre-printed with instruction and questions which take the team through a retrospective.
There are multiple sheets available for all of which follow the same basic format - similar to that outlined in Agile Retrospectives by Larsen and Derby. At the end of the sheet teams are asked to select three actions they will take to improve their development capabilities.
The layout and instructions on the sheet provide every participant with an opportunity to speak and facilitate the process. As a result there is no need to have a retrospective facilitator involved, the team can be self-organizing.
Some teams choose to keep a facilitator. One role a facilitator can help with is big teams (more than eight) need to split into sub-groups and use different sheets. At the end of the formal exercise the sub-groups compare conversations and notes.
Any stuation which requires a team to undertake a discussion, particularly a structured discussion can benefit from the dialogue sheet approach. Several teams have used the sheets for project kick-off. I use specially designed "Agile discussion" sheets at the end of training courses to promote reflection on course material and action planning.
Dialogue Sheets in Numbers
Downloads are free but the Dialogue Sheets download page askes downloaders for a few basis questions. I apologise to those who object to this process but having a little bit of data is very helpful and gives some very interesting insights both into Dialogue Sheets and retrospectives in general.
2012 download requests were a little short of 2,000, if you take out the duplicates and "Micky Mouse" registrations the number falls some more. Still, the number of valid downloads is more than enough to render these insignificant.
Downloaders claimed to be from 75 different countries. Perhaps unsurprisingly the greatest number were from the USA (a little over 18%). The UK comes second (a bit more than 13% downloads), again this is unsurprising given that I am UK based and most of the conferences and community groups I have taken the sheets to are in the UK.
French techies take third place with just over 7% of downloads. This does surprise me a little because the French have a reputation for preferring things in their own language. Credit here has to go to Laurent Carbonnaux who translated one of the sheets to French.
The rest of the top ten is made up of European and English language countries - Germany, Norway, Sweden, Australia, Canada and Holland. Squeezing in at tenth place is India accounting for just over 2.5% of downloads, below Holland but above Brazil. Given India's reputation for IT this might be a little surprising.
Location is interesting but I find the most interesting question to be "How frequently do you hold a retrospective?" Bear in mind that this question is asked when people download the sheets so it is indicative of the way the teams are, it tells us nothing about the effect of the sheets - that would be another survey I might run one day.
Over one third of downloaders report that they never hold a retrospective. Given the central role retrospectives have in Agile development this is both disappointing and sad. Add in the people who report they rarely hold retrospectives or only do one every six months and the number rises to nearly 45%.
It seems reasonable to assume that only those who are interested in retrospectives would visit the site and request a download. Thus it seems likely that the number of teams who actually undertake retrospectives on anything like a regular basis well below half.
More positively the number of people reporting that their teams do hold regular retrospectives (quarterly or more often) is fractionally below 50%.
Reports of use
The findings described in my previously piece continue to be consistently reported: all team members engage - even the quite ones; without a facilitator dynamics change, new issues are addressed and minority groups are more likely to have their voice heard.
I have been told multiple times that the sheets appear frivolous, or playful. For some people these are reasons not to use the sheets, for others they make them more intriguing. Undoubtedly there is a playful aspect to having a large, colourful, sheet of paper in a work environment. But then, who said work shouldn't be enjoyable?
Perhaps my favourite story to date concerns a team at Siemens split between the USA and India. The Scrum Master gave both locations a Dialogue Sheet and arranged for the teams to conduct retrospectives at almost the same time. She then gathered notes and findings from the two groups.
The two groups had discussed similar issues and come to similar conclusions. Although she noted that the Indian team seemed to be more thorough in their reasoning. This shows how the sheets can be used both for large teams and for distributed teams which often have problems holding retrospectives.
When I facilitate a sessions with multiple groups working on separate sheets I prefer to have each team present their findings. Once the sheets are complete - usually about an hour - each team in turn takes a few minutes to showcase their sheet. One team member will highlight the key points of discussion, read the concluding action items, and take questions from the larger group.
Another favourite story concerns a team in the North of England where the Manager reported "the difference from the last few [retrospectives] was like night and day." He went on to say no one held back and the whole team was actively engaged in talking and writing.
A Scrum Master in Norway reported that the sheets reduced retrospective administration significantly because findings were recorded as the exercise progressed and people could make any notes they wished on the sheet. Less bureaucracy meant more inspiration and a focus on the real issues.
She also reported a particular benefit for Scrum Masters who aim to foster self-organizing leaderless teams: the sheet rules prevent the Scrum Master falling into a "leadership-ish role."
As with several other teams this team hung the completed sheet on the wall after the retrospective to remind themselves of their findings and what they had decided to do. Another team reported returning to the hanging sheet during the sprint to guide conversations.
One team I heard from hung a new sheet on the wall at the start of the sprint. At anytime during the sprint team members can go to the sheet and add an event on the timeline or make some other note they wanted to consider during the retrospective. Thus the timeline is already fleshed out for examination at topics for discussion identified at the start of the retrospective.
Concerns and downsides
A concern of my own is that I don't hear enough reports of how Dialogue Sheet retrospectives went wrong or didn't work for a team. As well as a "too good to be true" feeling that can result it deprives me of opportunities to improve the sheets further.
That said there are some concerns and issues that have arisen. Some of these are specifically about the sheets while others say more about companies, projects and retrospectives in general.
One is that after using the sheets repeatedly teams become bored of them. In one case a team had adopted a mechanical approach to the sheets; sort of a "O, its that time of the sprint, we have to fill in the sheet again."
Dialogue Sheets should not replace all other retrospectives techniques. They are just another tool in the retrospective toolbox. As with any technique overuse is sure to lead to boredom. This is, in part, why I designed several variations on the sheets.
One of the big challenges with retrospectives in general is keeping them fresh and interesting. To offset this I advise teams to vary their approaches and try new exercises. Holding some retrospectives with facilitators and some without is one way of doing this.
An unexpected challenge was reported from a team in India: a company-mandated format forbid the team from changing the retrospective. When I first heard this I was flabbergasted. "If a team cannot change their retrospective what hope have they of changing anything else about the way they work?" I thought to myself.
I've had it pointed out since that maybe the company requires a consistent format to ensure consistent reporting. I can see some merit in this argument but if consistent reporting is the goal, cannot it not be achieved in some other way?
Unfortunately this is about as much as I know of this case. I'd like to understand this type of situation better.
Another concern, which is regularly reported, is the loss of the facilitator. Some teams feel they still want a facilitator in the loop. This shouldn't be a block to using sheets. Over 7% of downloaders actually describe themselves as professional facilitators. Just because Dialogue Sheets can be used to hold a facilitator-less retrospective does not mean they must be used to remover the facilitator.
I have personally observed several Dialogue Sheet retrospectives as a facilitator. Usually there is little for a facilitator to do in this scenario but for some teams it may be a useful safety valve. Indeed, a skilled facilitator may well use the Dialogue Sheet process as the start of a bigger retrospective. Teams seldom address all the issues they surface and a "fly on the wall" facilitator may note some of these to return to later.
An active facilitator may also address another issue that occasionally surfaces in comments: that the Dialogue Sheet does not create a very deep examinations of issues. To some degree this is a function of the time available to a team and potentially effects all retrospectives.
While one of the sheet designs does include a fishbone diagram for the team to use a facilitator may add this depth of inquiry that might otherwise be missing.
Another regular concern is that of team dynamics. Organizers report concern that the dynamics of the team do not lend themselves to a facilitator-less environment. There seem to be two main concerns: that some individuals may dominate the conversation, and that quieter team members may not say anything.
However these fears seem largely unfounded, one report stated: "Afterwards I got some very enthusiastic feedback from the people who I'd originally worried about."
Perhaps the most consistent finding from those who do try the sheets is that those who normally remain quiet, or have little to say, are more vocal and participate more fully when a Dialogue Sheet is used.
I suspect the fear is that without a facilitator to politely give others a voice one loudmouth will dominate the proceedings. Here I would recommend: keep the facilitator, let them set the retrospective going, then step back and observe. Only intervene if need be.
While one can never rule out someone dominating any conversation the design and rules of the sheets aim to give everyone a voice, thereby limiting the dominance of one person.
I continue to believe that one block, if not the major block, to using Dialogue Sheets is simply the need to print an A1 sheet of paper. One person did tape eight A4 sheets together but reported that it took him over an hour and he swore never to do it again! Taping four A3 sheets together is easier but still less than perfect.
A print-on-demand service does exist for the sheets, while the sheets themselves are not expensive postage can increase the price significantly. While I continue to explore options to bring the price down I am of the view that it is not the price of the sheet which is the block but rather the need to spend any money at all.
Considering the wage costs of having several expensive developers spend an hour doing a retrospective the need to spend less than $50 on a printed sheet is insignificant. However, spending any money at all requires authorization and this may be the real block. (Of course, this does lead to questions about how self-organizing a team might really be if they cannot spend even small sums of money.)
Workarounds exist, despite the hassle sheets can be taped together; some companies, particularly large ones, have oversized plotters available; and using a local print shop can be a quite cheap. Sheets can be reused with a little effort; sheets can be laminated or mounted behind glass in a cheap picture frame. These can then be written on with a dry-wipe pen. While this exclude the wall-hanging exercises mentioned above digital cameras can be used to preserve the findings.
With Dialogue Sheets are now established as an powerful retrospective technique the question is: What next?
I occasionally modify existing sheets to include new suggestions or observations and a new sheet was added a few months ago. This sheet takes the findings of a previous sheet as the starting point for the retrospective.
Translations are the obvious next step for Dialogue Sheets. Some have already been translated to French, Spanish and Finish and I hope to see Russian and German sheets this year. (If anyone is interested in translating the sheets or testing translated sheets please contact me.)
I am still interested in receiving feedback on Retrospective Dialogue sheets so if you have tried using them I'd love to hear your suggestions. And if you've not tried them I'd be interested to hear why not, is there something missing? Some obstacle I've not heard of before?
Cass Business School in London - from where I originally encountered Dialogue Sheets - also continues to experiment. They have triangular and circular sheets too, perhaps a circular on-going retrospective to perform root cause analysis would be possible? Or perhaps a circular, never ending sheet, to model an iteration, or the entire development process?
Other applications beckon. The game like appearance of the sheet often leads people to suggest a game of some sort. While this is an intriguing idea I've not yet been able to figure out the how make it work. Would it be possible to play a sort of Development-Monopoly? This would really change the nature of the dialogue but could result in powerful learning.
Right now I am experimenting with a sheet to guide new teams in planning meetings. Although by the nature of planning meetings such a sheet would involve less dialogue so it might be unfair to call this a Dialogue Sheet.
I'd love to hear of other ideas so if you have any please get in touch and let me know your thoughts.
About the Author
Allan Kelly has held just about every job in the software world, from system admin to development manager. Today he helps teams adopt and deepen Agile practices. He specialises in working with software product companies and aligning products and processes with company strategy. His latest book "Business Patterns for Software Developers" was published by Wiley earlier this year. More about Allan on his website and on Twitter as @allankellynet.