For a decade now, Kanban has offered an alternative path to agility. At the time of writing my first book, Agile Management for Software Engineering in 2002, I believed that an alternative path was required because too many people were resisting the adoption of Agile methods. After publication in 2003, I set about pursuing an alternative and Kanban is what emerged from that journey. So Kanban was for people and organizations who found Agile wasn’t a good fit for them, wasn’t suited to their culture or the nature of their business.
At the start of 2016, I’d like to take a closer look at why we need an alternative path to agility and identify the problems that are evident in the market with Agile methods and their adoption. The first in this series asks the question, “Is Agile costing you too much?”
In 2011, we collaborated on a conference in Finland called LESS. During a panel session which I facilitated, one question from the floor was “It seems to me, listening to the sessions at this conference that Kanban obviates the need for the Product Owner role in Scrum. I’d like to ask the panel, what would you do if you’d recently hired 400 product owners?”
Me, “I’m sorry, can I clarify that I heard you correctly, I thought I heard you say ‘400 product owners?'” “Yes, that is correct!”
“How many developers do you have?” “2,400. Our Agile consultants told us we needed one product owner for every six developers. So what should we do with these 400 new people we have hired?”
Last month, my colleague, Alexei Zheglov, was speaking with an Agile coach whose contract hadn’t been renewed. It is never nice to be laid off at Christmas. What had happened? Well, this coach responsible for coaching just six developers wasn’t renewed because there weren’t sufficient tangible benefits to justify the overhead. His contract was costing $150,000 every six months.
On December 24th, I met the CTO of a bank in China. This fairly mature IT organization was running a side-by-side comparison of Kanban against Scrum. The consultants and trainers for both methods were being supplied by the same firm in Shenzhen under a single purchase order. The Scrum coaching was requiring one fulltime Scrum coach for every twelve employees.
Late in 2014, I did some consulting for a large telecommunications equipment manufacturer here in the US. Specifically, I was advising a business unit based in sub-urban Boston. The general manager had brought me in because in his own words, “We’ve been doing Scrum for 6 years but we haven’t seen any improvements for at least the last 2 of those years.” The business unit had a Scrummaster for every 6 developers. I attended several retrospectives, usually facilitated by one Scrummaster and then each of the others in turn had an opportunity to speak up and report the Sprint for their teams. Later in the engagement I was discussing the need for someone to play the role of Service Delivery Manager for a kanban system. The GM asked me who should play that role. He suggested no one in his business unit had the skillset. So I innocently asked, “What do your Scrummasters do?” He replied that “they coach the practices of Scrum.” Since then I’ve come across many companies who describe their Scrummasters as Agile coaches and the ratio of coach to value-adding workers is in the range of 1:6 to 1:12.
The role of Product Owner is interestingly ambiguous in its definition. In some organizations, Product Owners are basically business analysts who write requirements. While in others, Product Owners have a role that is very specific to prioritization and they are the decision makers with respect to the sequencing of the work. Ken Schwaber described the Product Owner role as “the single throat to choke” and he clearly had decision making, responsibility and accountability in mind when he used that term. For many the Product Owner is an interface between multiple competing business stakeholders and the delivery organization (or Agile team.) The Product Owner is a curious addition for Agile. At the founding of the movement, Agile was in part about removing middle-men, eliminating signal attenuation and having direct contact with customers – conversations rather than documentation. It is strange, therefore, that to facilitate such a process we need a new form of middle-man, the Product Owner, to prioritize the order of the conversations. Regardless, of exactly what the Product Owner actually does in your organization, it does seem that often these product owners are new hires to the organization and again the ratio of 1:6 seems to be common. One of our own clients in 2015 made a decision to hire about 20 such Product Owners. While it goes against standard Kanban coaching guidance, the client had their own reasons and justification for the decision.
There is a question whether the product owner is a value-adding role or not? In many organizations it appears that it is not: The product owner is a middle-man with authority over prioritization. While sequencing and scheduling when done well will generate value, the act of sequencing one item over another, does not add value to the items themselves and from a Lean value stream mapping perspective it’s a non-value-adding role. Even more curious then that Agile requires you to create non-value-adding positions.
So in some extreme cases, Agile methods appear to have added 2 new people in non-value-adding positions for every 6 in value-adding positions. Put another way, after the Agile transition, 25% of the workforce is additional “Agile” overhead for operating the Agile method. While this is at the higher end of the range, I believe that it is typical for operating expense budgets for IT / product delivery teams to be increased by around 14% to cover the overhead of non-value-adding Agile facilitation positions.
What about Agile training? There are many 2-day, and in the case of scaled Agile solutions, 3 and 4-day training classes available. Many of these come with short examinations at the end and certifications for passing. However, what appears to be the case is that these training classes do not prepare the workers to behave in an Agile fashion or to deliver the promised benefits of Agile methods. The classes often contain abstract or metaphorical content and are filled with games which while fun to play do not teach specific actionable behavioral change in the workplace. All of this is okay, because the Agile transition purchase order includes coaching as well as training and the coaching is essential because the training did not contain anything actionable. The workforce, however, are now “certified”. Strangely this certification does not mean they know how to behave differently or how to produce better outcomes. It appears to mean that they are now certified to receive coaching.
Let’s imagine for a moment that Agile was producing twice as much value than what preceded it. If this were true would you spend 14% more on operating expense in order to generate that additional value? Rationally, you would say yes. It is an obvious and simple executive decision. The ROI is clear. And therein lies the rub with Agile. How many supposedly Agile organizations can show such benefits? The best and most popular Scaled Agile Framework appears to scale by offering quarterly release planning and using 3 month time boxes. The so called “release train planning” involves getting everyone in a room together in order to resolve dependencies between work and to devise a suitable schedule of six incremental 2-week sprints.
One company I know of in Central Europe, transports up to 500 people by plane, train or private cars, to a location near Frankfurt every quarter in order to conduct the scaled Agile release train planning. What is this costing? If we conservatively budget 800 Euros per person for transport, lodging, and catering, and assuming the company has a meeting facility large enough and isn’t renting a banquet hall from a nearby hotel, then we might have a budget of 320,000 – 400,000 euros per quarter.
Meanwhile, releases are only coming every quarter. How often were releases happening prior to the adoption of Scaled Agile? And how much functionality was getting released? Is there any tangible evidence that Agile is delivering the additional value it claims?
I hear other claims that Scaled Agile is leading to inefficiencies synchronizing teams with some teams having to take down time while others catch up. So the efficiency of Scaled Agile is being called into question by some of its practitioners. I haven’t seen this explicitly with any of our own clients, so it is hearsay, until I have some solid evidence.
Around 10 years ago I worked with a group of Agile coaches from Sabre Holdings, a firm which included properties such as Travelocity and Last Minute. They were under pressure to show an ROI for the Agile coaching team. The Agile coaching department was costing around $2,000,000 per year and the senior leadership wanted, quite reasonably, to know what they were getting for that money. After some period of time, the group was unable to show real tangible ROI and it was closed down. Some of those coaches found employment with other Agile firms such as Rally. After a period of time, Sabre was once again dabbling with Agile and the costs for consultants and staff augmentation were mounting up. A decision was made to create a new in-house team again. This time the justification was merely deferred cost against consultants and temporary staffing. It was no longer necessary to show an ROI on value created, just enough to show that the internal cost was less than using outside help.
So there have been 15 years of Agile and yet we still don’t have very many case studies showing tangible business benefits and real ROI. Meanwhile, Agile consulting firms thrive on supplying the large numbers of coaches required to “go Agile.” Perhaps, it is time to ask, “Is there an alternative approach to agility that costs a lot less?”
Back to the CTO of the Chinese bank that I met on 24th December. He told me that in 5 months, he’d managed to scale Kanban to 2000 people across 5 locations in China. That the adoption had institutionalized and he’d been able to make a tour of the various locations and inspect many of the Kanban implementations for himself. He told me that Kanban was producing greater productivity and efficiency than Scrum. He, however, didn’t elaborate much, and isn’t ready to share his data. He wants more evidence before he is prepared to go public. What he was willing to share is that the Kanban implementation had been achieved with a total of 200 days of training and consulting. This is a ratio of 1 day for every 10 employees. Meanwhile, Scrum coaching is costing 180 days for every 12 employees, in the same half year time period. He also confirmed that Scrum is failing to institutionalize and is requiring constant coaching.
To put this in perspective, in terms of per employee adoption cost, Kanban is costing this bank 1/150th of the cost of Scrum, and producing better results. How might this huge cost saving be possible? Well, for starters Kanban doesn’t require you to add people to your organization as coaches or product owners, instead it provides existing workers with a means to frame and make better quality decisions. Kanban is empowering and enabling. Workers become more effective at doing their existing jobs in the existing process and workflow – no need for coaching in a new role as part of a new process. It’s simply more effective to stick with what you do now and get better at it, than it is to make grand changes.
We are inviting executives from this Chinese bank to speak at our Spring conferences in London and San Diego. We hope to have both cost and value/productivity data to report by that time.
A week earlier I was in India where I saw a case study presented by a young consultant who was acting for a large Agile consulting firm, advising a large Indian outsource firm, to a large financial institution based in New York. Because of the nature of the contracts between these firms, the actual details are covered by a non-disclosure agreement, so we can’t name names, or any specifics. The client had adopted Scrum largely enterprise-wide and had adjusted all their processes to facilitate a 2-week Sprint cadence. Despite this highly Agile sounding arrangement, actual customer lead times for changes to credit card processing systems were taking 130 days on average. With some courage, the consultant took it open herself to push through a single kanban system implementation. With no other changes, within 6 weeks, lead times had shrunk to just 12.5 days and the client was so happy they are pushing the consulting firm to implement more Kanban. This particular consulting firm is well known to be anti-Kanban and their sales people around the world actively dissuade clients from adopting it. So it took a courageous act from a single professional keeping the client’s best interests at heart to create a huge improvement in value delivery.
That same week in India I saw another presentation where the kanban system had produced a 30% increase in revenue by enabling the firm to process more service requests faster. This is real ROI. Now would you spend 14% on OpEx to get 30% more revenue? Probably not! Revenue isn’t margin or profit. If you only got 30% more revenue but Agile was costing you 14% more OpEx you’d be losing money. The thing is, how many Agile implementations could claim 30% more revenue? Meanwhile, this Kanban implementation cost a handful of coaching days over 6 months. This story was a strange déjà vu with a case study presented at Lean Kanban Benelux by Tina Dudenhoeffer, a consultant, and Marianne Lanting of Winkle. Again, they had increased revenue by 30% in only 5 months by servicing requests for market research surveys faster and more predictably using Kanban. The cost of this implementation was less than 20 consulting days from Tina plus training from Maarten Hoppen of our partner VX Company. Marianne is a manager at Winkle and there was no additional cost for her.
In general these results while extreme sounding isn’t surprising. We already know from surveys and reports post Kanban training that our training is more effective than typical Agile training. Lean Kanban training is specifically experiential and uses games, simulations and exercises that simulate or model the actual working environment. Students leave our classes with specifically actionable ideas and designs for changes many of which can be implemented immediately at no or very low cost. They don’t need any or much coaching. They come out of the class understanding what to do and why they want to do it.
For 5 years now, we’ve been selling Kanban coaching engagements that typically look like 2-4 days per month for organizations up to a product unit (or 150 people) in size. Kanban coaching is typically centered around coaching managers on metrics, reporting, and holding the various Kanban Cadences meetings such as Service Delivery Review, Risk Review or Operations Review. We have often found ourselves bidding against Agile consulting firms recommending coaches are on-site for 4 or 5 days per week and often at the ratios we’ve discussed here 1:6 -> 1:12 people. Bizarrely clients have often opted for the latter, usually for 2 reasons. We don’t quote hourly rates for our consultants, only daily rates, and if you do try to calculate an hourly rate, it will be higher than that for a typical Agile coach. We aren’t ashamed of this. We believe our coaching is far more effective because it focuses on high leverage activities that deliver more value, rather than on Agile practice adoption and behavior. Meanwhile, the quantity of consulting days and the number of consultants is dramatically less. Often clients are skeptical of this. Everyone else is advising them they need a lot of Agile coaches and a lot of coaching. They believe it is too risky to go with our model when we appear to be bucking the trend. It seems easier to follow the herd. It is common for us to be quoting less than 1/10th the cost of rival Agile consulting firms, typically, $30,000 – $50,000 against $300,000 – $500,000 and then to be told we didn’t win the business because, and to quote a VP from a travel technology firm, “You are too expensive!” This analysis was based on an hourly rate and not on a total contract cost. Sigh!
So, while we have a bank in China that is rolling out Kanban at large scale at less than one-hundredth of the cost of Scrum, we have been and continue to be confident that Kanban, as your alternative path to agility, can be rolled out at only one-tenth the cost of competing Agile and Scaled Agile methods. We are prepared to put our money and reputation where our mouth is on this because we are more than confident with years of experience and real data that we can make it work. So, this month we are launching “The Kanban Challenge” – give us 10% of your Agile budget and we’ll promise to outperform the other 90% producing more and better results over a 6 to 18 month time period. Some conditions apply to this offer so do approach our sales team directly for full details if you are interested.
After 15 years of Agile, it is time that developing true business agility stopped costing you as much as 14% of your OpEx budget. So discover “the alternative path to agility” and take The Kanban Challenge in 2016. Let us show you how much is possible at just a fraction of the cost and impact of Agile methods. Watch out for the launch of The Kanban Challenge later this month. Until then please feel free to contact our sales team with any questions.