Behind the 2012 VersionOne State of Agile Survey
VersionOne recently released the results of their State of Agile Development Survey for 2012, and once again it proved to be an interesting indicator of Agile adoption and trends.
Scrum once again remained the most popular Agile method in use by 54% of respondents (compared to 52% in 2011):
54% Scrum 11% Scrum / XP Hybrid 9% Custom Hybrid 7% Scrumban 4% Kanban 4% Don't Know 11% Other (including XP 2%)
The most popular Agile techniques in use remained the same as previous years, although almost all of them saw a reasonable percentage increase since the 2011 survey
85% Daily Standup 75% Iteration Planning 74% Unit Testing 72% Retrospectives 69% Release Planning 67% Burndown / Team-Based Estimation 58% Velocity 57% Coding Standards 56% Continuous Integration
When asked about the leading causes of Agile failures, 18% of respondents reported that none of their projects failed, while those that had failures reported issues like company philosophy or culture at odds with core agile values (12%), external pressure to follow traditional waterfall processes (11%) and a broader organizational or communications problem (11%). This year the survey dug into the organisational issues behind failures and found the following:
34% Failure to integrate people 28% Failure to teach team-based culture 21% Communication between Dev / Product-Owner 9% Communication between Dev / QA 8% Scrum Master problem
The survey also showed that overall Agile momentum is rising with those with plans to implement Agile up from 59% to 83% this year. Interestingly though, fewer people are planning to use Agile on outsourced projects falling to 49% from 77% last year.
With so much data to interpret, InfoQ caught up with Robert Holler, President and CEO of VersionOne, to talk about the results of the State of Agile Development survey.
InfoQ: What is the driver for Version One conducting this survey?
As part of our marketing and sales process, we are constantly asked questions around the market and adoption and there is not a lot of insight into that, especially not aggregated into one survey. The survey started seven years ago and it has grown and become recognised and anticipated each year. We get a lot of insight out of it and we hope it helps the industry understand what it going on and what things have changed. Each year we always see an upward tick and positive trend but this year we have seen quite a change in momentum, for example in relation to transition last year it was 59% either doing Agile or planning to do Agile to 83% this year, it just doesn’t get any more significant than that. In terms of saturation, the number of folks with more than one year experience going from 74% to 81% and the number of people with 5+ years experience going up to a quarter of the people who answered the survey. Agile is finally getting to the point where we believe it is poised to explode, particularly in terms of scale in large organisations.
InfoQ: A survey such as this relies on people being knowledgeable about Agile or Version One in the first place, in that case would you consider this survey to be more of a State of Agile Community survey rather than just a State of Agile survey?
Yes, it is more a state of Agile community, 80%+ are doing Agile but you do have 10% to 15% who aren’t but it is clearly not the majority.
InfoQ: You mentioned earlier about momentum and experience, are there any other learnings that took your attention this year?
One of the new questions we asked was about criticality of executive support, and the question really revolved around when you are scaling Agile beyond a team, what are the most critical considerations, and the fact that immediately bubbled up to the top was interesting to me although you contrast that with where the least amount of Agile knowledge is it just points to the fact we have some work to do there in terms of education and successful adoption. What we have really seen there is that executive management is typically going to come around when success has occurred such as reduced risk, accelerated time to market and improved quality and those are the things we are actually seeing going on right now in droves.
InfoQ: Does the survey reflect what VersionOne is seeing from its customers and the community in general?
Absolutely. The shift we have seen in the last couple of years in particular is Agile extending outside of the team to more of a division or enterprise level phenomenon and when you look at the survey results of 5+ teams and 10+ projects there is a lot of evidence to say there that this thing is starting to scale and starting to scale significantly and its been successful doing so. I’d say it very closely matched our experience of sales, marketing and development.
InfoQ: Looking at the survey itself, in the results of Who Knows Agile, it is interesting to see that there is a low knowledge in the non-technical areas like QA, Executives and Product Owners. What do you think we can learn from this?
I would tend to agree that probably the most surprised I was was around the Product Owners. I work in a software company and I tend to work around other software companies and the Product Owner in our environment is probably one of the most knowledgeable. We have however, especially over the last few years, transitioned out of selling primarily to software companies and much more into large enterprise or service industries such as manufacturing, telecoms and government and it was starting to become obvious to us where there was a fall off in terms of education. Being a Product Owner is not an easy thing to do, especially in a large organisation and they are probably pulled in seventy different directions and being given the authority and the empowerment to manage the backlog and get it to the right level of granularity and groom it on an ongoing basis is not that simple of a thing and I think that kind of manifested itself in the results. It points to the fact the development team either get it or think they are getting it but outside the development team there is still a lack of knowledge.
InfoQ: The survey shows that 63% of initial champions of Agile methods are management (including executives it is 76%). Do you think this is this helping or hindering the adoption of Agile?
About 13% were executives. The rollup in the survey in relation to management is going to be project managers and Scrum Masters and is a broad array of people who manage the software development process. This where most of the initial champions going back through the last several years have come from, whereas if we go back to the early days of the survey it was really coming out of the development team. The management level is where the interface and translation occurs and that is where the knowledge resides (at least that is what the survey says), so that is where the initial champions are and that points to good progress in terms of adoption.
InfoQ: In relation to the number of projects using Agile, 59% of respondents say they have 0-5 projects. Is this number telling us that only a handful of projects in a company are Agile or that many companies have a small amount of projects?
We obviously need to do some deeper analysis here, but the good thing out of that exact question was from a momentum standpoint the transition was companies that had more than 5 projects going on. Last year it was 29% and this year 41%, so that was kind of a significant change. In relation to the less than five it still points to the fact that most software development, at least in terms of numbers of projects, happens on a small basis. You have tens of thousands of organisations out there that do software development, you don’t have nearly that many that do software development at significant scale. So just in order of numbers, it is multiple orders of magnitude of organisations that are just doing Agile on a small scale. My guess is a majority of those are smaller organisations.
InfoQ: I guess that reflects in the number of teams, 30% have 10+ teams and 38% have 2-5 teams.
Yes, the total median size of organisations is around 100.
InfoQ: The distributed numbers surprised me. Is there anything we can read into the data that only 35% of respondents are doing distributed development? Was it indicative of certain types or locations of companies?
When we asked that question we didn’t just ask a yes or no question, we asked an open ended question of how many teams do you have doing something, and you never know how people are going to answer a question like that. In future years we will probably simply ask do you have distributed teams doing Agile, with a simple yes or no. I have a feeling people may have got confused, because in our experience that does not quite reflect what we are seeing. We see easily half the companies we are dealing with, even small ones, have distributed team members. What does distributed teams mean – different locations, different sites or different cities? We’ll clarify the question better next year, but I was very surprised at that one too. That said, one of the reasons our customer base picks our tools is because it supports distributed teams, so our universe may be a little biased, but in terms of other surveys I have seen and other community activity going on, the majority of them are easily doing distributed development of some sort.
InfoQ: Is the fact that 34% of respondents do not know what APM is indicative that many teams are stopping at the basics of Scrum deployment (also indicated by the large amount of Scrum in use in the survey)?
If we go back 18 months, I would go as far to say that no one even knew what APM was so I think that is a natural progression of knowledge. I think as coaches we have a natural tendency to stay on the front edge of these things whereas the mainstream market doesn’t get anywhere near to that. I looked at it as at least a quarter of people knew what it was! I thought that was actually impressive because I think 18 months ago that it would have been less that 5%!
InfoQ: So what you are saying is that you are glass half full and I am glass half empty?
Yes! We are just so entrenched in Agile, it is just second nature for us to know, we have to know about Agile and the latest practices. Look at Kanban, not that many are practicing that, yet it has been out in the software development community for at least the last five years. I think APM and Enterprise Agile is headed down a natural progression, we are only a year or two into it, but the community who is answering these questions is a fair representation of the actual community and they are not paid to stay on the cutting edge of Agile.
InfoQ: In relation to that, I guess it is not surprising that the survey shows that 54% of respondents are using Scrum if you take into account the number of large enterprises and the large user base in North America and Europe. The 4% of Kanban is low, although you noted in your comments the variance almost doubled this year?
One thing to pay close attention to, while Kanban as a primary methodology may only have less than 5% as a technique used, if you actually look on the same page of the survey, it is actually one of the faster growing techniques with a third of companies practicing it (up from a quarter last year). Whether that means a Kanban board or work in process or whatever, it may not be their primary methodology but the techniques are absolutely being used in much greater numbers than just the primary method.
InfoQ: The results of the role of the Scrum Master are interesting. Was there anything in the data that was consistent amongst those who actually had that title when responding to the survey?
Not in the data directly, but in our world we are seeing that folks are trying to time slice for their managers. Scrum Masters are the new new Project Manager and are still trying to be time sliced. I see that is a significant difficultly but it is not just Scrum Masters, it is project team members as well as in our world we see that people are still working on three projects simultaneously. We obviously would prefer that not to be the case but like you recognise that is not the case, but you get into larger and larger organisations and you see that is just the norm. However, I have seen a trend that folks are fighting against that; it may have been the norm three years ago but now you have a lot of teams that are stepping up and saying “we don’t have the capacity to work on that, if we want to get this project done we need to be a little bit more serialised and we can get it done faster”. I think there has been a bit of a backlash recently away from that, but it does not mean it is not going on in a high percentage of places. You do see a lot of two and three projects, not a lot of five, six, eight, ten projects but very often in the two or three range.
InfoQ: 52% of respondents still struggle with changing organisational culture and this has been a theme in the last few surveys. Do you think these barriers are good indicators as to where more work is required in the Agile community?
When people unfortunately treat Scrum and Kanban as an automation or a mechanical process it just does not do it justice. If you are not treating it as a change management scenario then you are absolutely making a mistake and I think that mistake is made quite often by folks who do not fully understand what change management is. I think it points to if you are going to change the way you are doing software development, you have to approach it as two separate projects. One is I have a project going on to deliver something and I’ve also got a project going on to change my organisation. I think it is wrapped up into two things; cultural change and natural resistance to change and I think those things are highly interrelated. Cultural points to operating not in a command and control environment (those things just take time) and the second is a human element and humans don’t change overnight that easily. Those two have been at the forefront for many years, and I don’t see it changing anytime soon. When you start talking about fundamental change, it doesn’t matter if it is software development or other, there is going to be a natural resistance to change and something like this where you are not only doing things technically or mechanically different but philosophically you are doing things different, there’s an impediment to change or inertia that has to be overcome and that inertia is just typically overcome, over time. You cannot sometimes force cultural change, it just has to happen over a progression of time.
InfoQ: The satisfaction with VersionOne in the survey as a tool was above all other tools. Why do you think that is?
Every year we ask the question around tools and we know that we are the ones who distribute it and we work very very hard to get it out to as many channels as humanly possible. We try to take out some of the bias there, but clearly we are also sending it to our contact database so there is obviously some potential bias there. What we try to do is think about what is the other question that we could ask that’s normalised that would take that bias out of it and take another view that isn’t related to the number of respondents. Honestly, when we asked that question we had no idea what the answer was going to be, we took a little bit of a risk there and thankfully it worked to our favour. Do I have a full explanation for it? First and foremost I was just pleased, secondly I wonder what exactly all the reasoning behind it is. The way that we go about promoting and selling our product, we try hard to do it in context of a community commitment, we do listen to our customers, we are aggressive about trying to solve some of the more significant issues and we have quite a community of fairly strong followers. We are smaller, we are little bit more nimble, we are incredibly responsive from a customer support, sales and marketing standpoint and typically we get a lot of positive feedback along those lines. So I think we have a reasonable perception generally, and there are others in the list like that. LeanKit had a wonderful response for their customers, Atlassian with JIRA always gets a positive response from their customers, so we were glad to be put in that category.
InfoQ: The If you could say one thing... question is new. Were there any good responses that you didn't publish?
We highlighted what we thought was important but what I got out from reading many of the responses was that fundamental to Agile is education and training. If you are not going to provide it then why do it. If you don’t have a knowledge base and if you don’t have a coach you are hampering the ability to succeed. The other interesting one in a number of responses was that Agile is not going to be an overnight success, you have got to give it time. There was numerous examples of you have to give it time, you have to have a champion and you have to drive it through the organisation. Not that you can’t deliver some success in the short term, but if you are going to scale you have to give it the time it is due. It is cultural change, it is organisational change, it is philosophical change and it is metrics and process change. It just takes a while to permeate the organisation and a lot of the responses revolved around just simply giving it its due in terms of time and commitment.
InfoQ: What has VersionOne learnt from this survey that you can apply to your product?
We asked a question slightly differently this year. Instead of what tool do you use it was more how many or what number of and it was amazing! Some people use 15 different tools, but the average was 3-4 different tools. That is continuous integration and the like but it points to the fact that while we might be a central hub from a project management standpoint we clearly can’t assume that we are operating independently. We are operating interdependently with source code, continuous integration and build tools and deployment tools and we need to make sure that our product works in well and plays nicely. We are actually up to 45 different integrations now, but we can always do a better job of making things easier, making things less painful and improving the level of integration we have with other tools. The lesson that I got was that you are always in a situation where you have got multiple tools and therefore the onus is on us to work and play well with others.
InfoQ: Unrelated to the survey, but what exciting developments are coming up from VersionOne in the near future?
Two of the most exciting things that we have had the initial releases of and we continue to enhance are firstly interesting visualisations like bubble charts, project dependencies and amorphic tree diagrams. When you have got lots of work items flowing through a software development cycle and they are changing and there is interdependencies, wouldn’t it be great if I could visualise that stuff a little better. So we launched something at the end of 2012 and we continue to get wonderful feedback on that. It was not something that people were asking for in droves, but when you have a change oriented project management philosophy and it is happening at hyperspeed it is easy to lose track of all the dependencies and interdependencies amongst work items and how things roll up and how things are aligned, so we have put some really nice visualisations on top of that. Secondly, we launched a team room which basically strips out all of the management overhead out of the product and delivers a separate UI just for development teams that allows us to focus just on the daily activities and the communication and collaboration around that and changes to items that they are focussed on. That has been a huge adoption exercise for us to get teams to appreciate and adopt and roll out Agile and our product that much faster. More recently we are focussing on customer and team collaboration – there are some pretty exciting things coming up there. We are also looking at a next generation of our API to allow people to consume and build integrations that much easier. Definitely we are also continuing down the enterprise Agile and APM path – it is significant inside of enterprises, they want to scale Agile, they want to share alignment between strategy and execution and we are absolutely committed to that.
The full results are available from the VersionOne website. You can also look at the InfoQ summary from the 2011 survey. Are there any results in this years survey that are surprising or go against the state of the Agile community?