LeSS Huge Framework
• Requirement Areas •
With 1000 or even just 100 people on one product, divide-and-conquer seems unavoidable because of the complexity of so many requirements and people. Traditional large-scale development divides these ways:
single-function groups (analysis group, test group, ...)
architectural-component groups (UI-layer group, server-side group, data-access component group, ...)
This organizational design yields slow inflexible development with (1) high levels of waste (inventory, work-in-progress, handoff, information scatter, ...), (2) long-delayed ROI, (3) complex planning and coordination, (4) more overhead management, and (5) weak feedback and learning. And it is organized inward around single-skills, architecture, and management, rather than outward around customer value.
But in the LeSS Huge framework when above about eight teams, division is around major areas of customer concerns called Requirement Areas. This reflects the customer-centric LeSS principle.
Size—A Requirement Area is big, usually with between four and eight teams, not one or two. The following Area Feature Teams section on p. 35 explains why.
Dynamic—Requirement Areas are dynamic. Over time an area will change in importance, and then it grows or shrinks with teams joining or departing—most likely to or from another existing area.
Example—For example, in a Securities product (to trade stocks), these could be some major areas of customer interest—Requirement Areas:
trade processing (from pricing to capture to settlement)
asset servicing (e.g. handling a stock split, dividends)
new market onboarding (e.g. Nigeria)
Conceptually in the one Product Backlog, a Requirement Area attribute is added, and each item is classified into one and only one area:
Then people can focus on one Area Product Backlog (conceptually, a view onto one Product Backlog), such as the market onboarding area:
Common Sprint—Does each Requirement Area work separately in its own Sprint, with delayed integration until a far-future date? No.
• Area Product Owners •
In LeSS Huge one new role is introduced. Each Requirement Area has an Area Product Owner who specializes in that area and focuses on its Area Product Backlog.
Large product groups usually have several supporting product managers specializing in different customer areas, and some of these are likely to serve as the Area Product Owners. Sometimes the Product Owner also serves double duty as an Area Product Owner for one area; that’s more likely in small less huge LeSS Huge groups!
• Area Feature Teams •
Area feature teams work within one Requirement Area (e.g. asset servicing), with one Area Product Owner focusing on the items in one Area Product Backlog. From a team’s perspective, working in the area is like working in the smaller LeSS framework—they interact with their Area Product Owner as though she were the Product Owner, and so on.
The team members come to know the customer domain of that area well. And fortunately, the items of one Requirement Area tend to cover a semi-predictable subset of the entire code base, thereby reducing the scope of what they have to learn well within a vast product.
Key point about size: Many feature teams work in a Requirement Area.
The Magic Number Four
First, why does a Requirement Area have a suggested upper limit of eight teams? See The Magic Number Eight, p. 12.
What about the lower limit of four teams? Why not one or two teams? Naturally, four isn’t a magic number, but it strikes a balance so that the product group is not composed of many tiny Requirement Areas.
What’s the problem with many tiny areas? They reduce visibility into overall product-level priorities, increase local optimizations, increase coordination complexity, require more positions, and create teams that are too narrowly specialized and lack the flexibility (agility) to take on the emerging highest-value items from a company perspective. Furthermore, in a tiny area the Area Product Owner is increasingly likely to act as a business analyst between the users and one or two teams.
Are there any reasonable exceptions to the lower limit of four? Yes:
An early transitional situation when the group is incrementally growing a new area that is fully expected to ultimately have four or more teams. Then, start small and simple with one team.
When re-balancing teams from an area with a decreasing demand to one with an increasing demand causes an area to go from four to three teams. Ultimately, merge two reduced small areas back into a new larger area.
Example Requirement Areas and Teams
In summary, a Securities product could have
one Product Owner and three Area Product Owners, all together forming the Product Owner Team
six feature teams in the trade processing area
four feature teams in the market onboarding area
four feature teams in the asset servicing area
• LeSS Huge Framework Summary •
Each Requirement Area works as a (smaller framework) LeSS implementation, each working in parallel in one overall Sprint. We sometimes summarize a Sprint in LeSS Huge as a stack of LeSS.
As with LeSS, there are rules and optional guides for LeSS Huge; those are introduced in the following stories and fleshed out in later chapters.
Roles—Same as LeSS, plus two or more Area Product Owners, and four to eight Teams in each Requirement Area. The one Product Owner (who focuses on overall product optimization) and the several Area Product Owners form the Product Owner Team.
Artifacts—Same as LeSS, plus a Requirement Area attribute in the one Product Backlog and thus an Area Product Backlog view for each area.
Events—There is still only one common Sprint for the product; it includes all the teams and ends in a common potentially shippable product increment.
• LeSS Huge Stories •
Learning LeSS Huge—Readers who prefer exposition can comfortably skip ahead to following chapters, bypassing these stories.
Simple stories—These are intentionally plain and simple stories just to introduce basics in LeSS Huge.
Two topics—Following are two stories with distinct topics:
Creating and growing a new Requirement Area to deal with a new gigantic requirement.
Working with multi-site teams. (This happens in the smaller LeSS framework too, but is especially common in LeSS Huge.)
• LeSS Huge Story: A New Requirement Area •
Priti welcomes Portia to her first day in her new job.1 As a mid-level Operations manager in the Securities division of the large trading company as well as Product Owner for their internal Securities system, Priti is also responsible for finding and retaining talent for her Product Owner Team of Area Product Owners. And she thinks Portia is a fantastic find, as her expertise is exactly what is required for dealing with some new huge requirements.
During the recent job interview—when Portia was still a product manager specializing in regulatory issues at a company that made a system for trading bonds—Priti had laid out the situation. “Portia, after the last crash, the regulators are coming down hard and they require us to be compliant with Dodd-Frank. Right now, we don’t know what it exactly means or how it will impact our system. You’ve got incredible knowledge of this space, and a great professional network with the regulators. I would love it if you would join our group and help us figure out how to deal with this.”
A Big Surprise
A few days later... Priti welcomes Portia, Peter, and Susan into her office. Peter is Area Product Owner for market onboarding, and Susan is a Scrum Master from the trade processing area.
Priti says, “As you know, Dodd-Frank is coming, and it’s huge. What you don’t know is that this morning the regulators called us and they want us to take action now. I’d been working under the assumption we could start next year. So we’re going to have to adapt, big time.
“I don’t think anyone is clear what it means in detail—even the regulators. And we don’t know how it will impact our system and how much work this is going to take, other than, a lot! But now Portia’s joined us and she has a better understanding of this than anyone, although she’s totally new to our systems. So, how can we help her start tackling this mountain of work?”
Susan asks, “You guys understand the Dyslexic Zombies, right?”
Peter and Priti nod. Everyone knows about them—and it isn’t just their name. The Dyslexic Zombies1 have probably the broadest experience of all the teams. They’ve been around for years and they were a true pain in the ass when they adopted LeSS. The team contained two former members of their now-abandoned architecture group and a couple of people who had been working on the system for over fifteen years. Those people’s resistance to the LeSS adoption was legendary as they were afraid they’d lose their “system perspective.” To their surprise, the opposite happened! Because of their deep knowledge they continuously get tough items to develop. And they regularly participate as expert-teachers in current-architecture-learning workshops with newcomers, and Mario—one of the former PowerPoint architects—is now coordinator for the architecture community. When fed enough beer, he’ll admit that working closer with code and tests has increased his real understanding of the system.
Susan continues, “If any team can quickly help Portia get a better understanding of the size and impact of Dodd-Frank, it’ll be the Zombies. And they led the work on Sarbanes-Oxley a few years ago. Tomorrow is their PBR session. They are just about wrapped up on a new feature. Why don’t we re-direct the meeting to include them in a discussion on Dodd-Frank, and soon after, ask them to focus full-time on it?”
Refining with Zombies
Next day at the refinement meeting with the Zombies, Portia explains the situation, “You’ve probably all heard about the Dodd-Frank legislation. But here’s the surprise: We’ve just been told by the regulators that they want us to take action ‘now’ and demonstrate significant compliance by the end of the year. Otherwise they might restrict our trading.”
The Zombies are visibly surprised. They had heard rumors but didn’t expect such a rush!
Mario says, “OK Portia, give us a quick summary of what this means. And how is it different from Sarbanes-Oxley?”
Portia picks up a pen and starts sketching on a whiteboard. After about 45 minutes, she is finished with the overview and the Zombies looked a little stunned.
“End of the year, they said?” says Mario. “If the whole group started today, it wouldn’t get finished. This is huge!”
He takes a pen and at the whiteboard starts a rough sketch of their system, talking with the other Zombies about the impact it might have.
He says, “Portia, let’s also use this as a chance to help you understand the system better. Ask away.”
Portia says, “Can you hold on for a second? Let me start a video recording to help me remember this.”
Michelle, a veteran in the team, says, “We’d better start on some real development soon and learn more as we go because otherwise we’ll end up analyzing forever. I’ve seen this story before.”
Susan, their Scrum Master, says, “Reminds me... Tom DeMarco once said that the reason for every failed project is that it started too late.” Everyone laughs. She continues, “So here’s a suggestion: take a bite.”
Creating a New Requirement Area
The next day, Portia, Priti, and rest of the Product Owner Team meet. Portia shares a summary of the scope as she understands it now.
Priti says, “This is even bigger than I expected, and we need to show some tangible progress to the regulators within a few months, and major progress before fiscal year end—seven months from now. To state the obvious, they’re now authorized to require more from us, and with the power to shut us down. As you know, just last month the CEO made it crystal clear that new regulatory requests take priority over any other concern. It’s my experience that our goodwill and flexibility with the regulators goes up if we can give them something early, and be transparent and responsive. So that’s what we’re going to do.”
Priti continues, “It seems to me that we’ll need a new area for this big surprise. And of course that’s probably going to impact some of our existing high-priority goals, since we’ll have to shift some teams. Let’s prepare for a deeper discussion of overall prioritization impact in a couple of days. But for now, I’d like your input about spinning up a new area.”
After a short discussion, it’s clear that everyone recognizes the importance of creating a new area.
Priti then says, “Portia, I know you are new to us, but do you think you would be able to handle the Area Product Owner responsibility for this?”
Priti continues, “Peter, do you think the Zombies could start work on this? And we’ll need them to learn more Dodd-Frank and figure out the impact on our system before we can add more teams to this.”
Peter says, “I don’t think we’ve got any choice.”
Priti says, “OK Portia, so currently we’ve got a few items in Peter’s Area Backlog, the one huge item I think you called “remainder of Dodd-Frank” and the tiny item which the Zombies and you split off of it. Please ask Peter to show you how to set up a new area in the Product Backlog and move the items over to it.”
Priti continues addressing the group, “The next Sprint starts in three days. Let’s move the Zombies into your area and get started on this monster. Probably in a couple of Sprints we’ll be ready to—and need to—grow your area by moving in another team. Folks, please think about two major concerns: First, preparing for a serious prioritization impact meeting in a few days. And second, what other teams will be good candidates for the new area.”
Sprint Planning in the New Requirement Area
Each Requirement Area holds its own Sprint Planning meetings, all more or less in parallel. In Portia’s new area, she starts her Sprint Planning by introducing two unfamiliar faces to the Zombies.
She says, “Gillian and Zak have been in contact with the regulators regularly and will help us flesh this thing out. They’ve agreed to help us now in Planning, during our PBR sessions, and as much as they can spare daily during upcoming Sprints.”
She continues, “Here’s my tentative plan of attack for the next two Sprints. First, together we need to learn more about Dodd-Frank, and also split it into some major and manageable pieces so we can start to clear the fog and get a better sense of priorities.
“Second, we implement the smaller bite we’ve taken, starting this Sprint. That’ll give us better information about the real work and the impact on our product. And we’ll have some concrete visible progress.
“Third, we prepare for more teams to join our area. What do you think of this approach? Other suggestions?”
During the short discussion, Mario says to his team, “Let me give a bit more context, because I represented our team in the recent Product Owner Team meeting with all the Area Product Owners and Priti. To start with, it’s just us to start. We’re going to take the lead on early implementation, and getting the big picture of the item, and understanding the overall impact on our architecture.”
Michelle interrupts, “Like a tiger team working on a new product?”
“Yes, like that,” says Mario. “Think of Dodd-Frank support as a new product that needs to be continuously integrated into the rest of the product. But we’re in a hurry and it’s a ton of work, so in a few Sprints one more team will join us and shortly after, probably two more teams. We keep developing too, but we’ll be the leading team, which means we’ll need to bring the other teams up to speed and make sure we keep the overall product in mind.”
Michelle says, “It’s starting to sound to me like we’re going to become the architecture and project management team!”
Mario laughs, “No. I’m done with that. We’re still a normal feature team, but besides development we’ll focus on mentoring and bringing the new teams up to speed as fast as possible. But let’s be clear: team coordination and management is still the responsibility of each team.”
The First Sprint in the New Requirement Area
Their first Sprint is an unusual balance of clarification versus development, but nevertheless quite useful in this extreme situation. They spend almost half the Sprint in clarification with Portia, Gillian, and Zak. That’s because even for this extremely small bite, trying to understand what is wanted in the obscure realm of new government regulations—with no direct access to the politicians and policy writers—required a lot of investigation, reading, discussion, and communicating with outsiders. They expect that in future Sprints, the amount of time needed for clarification will soon drop down to a more common 10% or 15% of their Sprint.
And so they also only spend about half the Sprint developing one small item. But the discussion and the learning from coding pays off. Slowly but surely they start to split Dodd-Frank apart—at least the parts that any of them can understand.
While implementing the small item they had bitten off first, they spend much of the time together at whiteboards to discuss the overall design implications on the system. The team moves frequently back and forth between the code and the wall.
Sprint Review in the New Requirement Area
The overall Securities product group works together in one Sprint, with one final shippable product increment. But each Requirement Area holds its own Sprint Review, all more or less in parallel.
In Portia’s area, during their Review, she, Gillian, and Zak explore the one “done” item that the Zombies have managed to complete and integrate into the overall product. They had originally forecast two items, but Portia is impressed that they got even one done, given how fast this new work was thrown at them.
The Second Sprint
In the second Sprint they’re able to make slightly better progress on items, though they once again spend a lot of time clarifying together with Portia, Gillian, and Zak.
In the middle of the Sprint they hold a multi-team PBR session with the second team that is planned to soon join the area, teaching them about Dodd-Frank. They hold a current-architecture learning workshop to introduce the team to the major design elements already in place.
The Zombies know how big the work is and look forward to more help.
Product Owner Team Meeting
A few Sprints later... It’s time once more for the per-Sprint Product Owner Team meeting. They use it to align and coordinate between the different Area Product Owners, and for Priti to give guidance.
The Area Product Owners each share in turn their situation and upcoming goals. When it’s her turn, Portia says, “To none of our surprise, the progress is little and the surprises are big. But the fog is clearing and the teams and I are getting our heads around the work. Gillian and Zak have been tremendous help.”
Pablo, the Area Product Owner of asset servicing, comments on some close item relationships he now sees between their areas. Portia agrees to meet with Pablo and some team representatives later.
Priti asks, “Portia, about our upcoming Sprint. What are your goals?”
Adding a Third Team
Two Sprints later... At the Product Owner Team coordination meeting, Priti says, “As you know, Portia’s area still has only two teams. I know that Pablo would like to keep his six teams in asset servicing, but Dodd-Frank is just too important to me this year. So we’re going to move one team from Pablo’s area into Portia’s. Pablo, please ask for a volunteer team from your group and let me and Portia know.”
Some key points from the story in LeSS Huge:
The Product Owner is responsible for finding Area Product Owners and developing their talents.
The Product Owner is responsible for deciding to start, grow, or wind down Requirement Areas.
Requirement Areas are large, normally requiring four to eight teams, but during initial startup they may be smaller, especially if initiated with one team using a Take a Bite approach.
A Leading Team works solo to tackle a gigantic item until they understand the domain and development, and then they coach more incoming teams to help with the vast work.
• Multi-Site Teams: Terms & Tips •
Next is a LeSS Huge story involving multi-site teams. But first, some clarifying definitions, because the common term distributed teams confusingly means several things. The clarifying terms are as follows:
dispersed team—One team of (e.g. seven) people spread out in different locations; either different rooms, buildings, or cities
co-located team—One team working literally at the same table
multi-site teams—One co-located team working at one site, and another co-located team working at another site
Second, an observation and guidance:
A dispersed team is rarely a real team; it is much more likely a loosely connected groups of individuals. The communication and coordination frictions are higher, and they seldom jell as a team.
When your product group is 50 or 500 people, dispersed teams aren’t necessary. Each team of seven-ish people can easily be co-located. However, some teams may be in different sites, so that the product group has multi-site teams. Dispersed teams are usually the result of bad organizational decisions and ignorance about the cost of not having co-located teams.
• LeSS Huge Story: Multi-Site Teams •
Portia is the Area Product Owner for a new Requirement Area in a Securities trading system. The new area started with just one team for focus and simplicity. A few Sprints later Portia’s area adds a third team. Her first two teams are based in London with her. But her third new team, HouseDraculesti, is based in Cluj Romania at a major development site for the company.
Why not add a third team from the London site? That would have avoided the many aggravations and efficiency penalties that can come from multi-site development within one area—costs potentially so high that adding a team can effectively result in deleting a team.
But on the positive side in this case, Cluj is only two time zones from London, and everyone there speaks English well. And they are all strong developers with Computer Science degrees, in a city that values long-term and hands-on engineering mastery. Also, this is a dedicated internal development site for the company, so these are experienced internal teams that have in-depth knowledge of the product and domain.
And bottom line, Priti (the Product Owner) didn’t want any of the other London teams to shift from their current areas.
Priti knows that multi-site teams are a new situation for Portia, and so at their next meeting, she says, “Please ask your Scrum Master to talk with Sita, and also ask Sita to coach some of your events. She’s a Scrum Master in asset servicing, and she’s observed their multi-site situation for a few years. She knows the importance of Scrum Masters co-located with their teams, and she’s helped facilitate many multi-site meetings.”
Priti continued, “Also, we’ve had a super profitable year, so I’m providing funding for you and the Zombies team—at least those that can travel—to spend a Sprint in Cluj as soon as possible. Work closely with them, all in one room. The Cluj team could come here to London, but you want to send a strong signal that they are important, at their site. Try to avoid making them feel that London is more important than Cluj. Oh—and you’ll want to regularly visit every few months.”
Multi-Site Sprint Planning Part One
A few Sprints later, Portia walks into the room. There’s a computer projector attached to a laptop, displaying via video a room in Cluj. The whole team in Cluj are sitting and waiting. Sita suggested it would improve learning and engagement if the entire Cluj team participated in multi-site meetings for the first few months of their addition to the area.
All the team representatives have tablets or laptops with them.
Portia begins. “Welcome and let’s get started. My offer of items this Sprint are highlighted in the shared spreadsheet. Can you all see it? I think you all understand why these are the themes and priorities, since we’ve been discussing this in PBR and it reflects your input and mine. But please ask again if you’d like clarification. Other than that, you’re invited to enter your team names beside the items you want.”
That done, the group enters a Q&A phase to wrap up lingering questions about the items. The London representatives tape up some flip-chart papers and start writing questions. The Cluj team members enter their questions in separate sheets of a shared spreadsheet. Portia spends some time at the different paper flip charts, discussing answers and sketching on the paper. And she spends some time at the spreadsheet, typing in answers for the Cluj team, while also talking with them face-to-face via the video session.
After about 30 minutes the separate questions have been resolved, and Portia asks everyone to come back together. She says, “Any issues or questions that you want to discuss together, before we wrap up?”
Multi-Site Overall PBR
People enter the workshop room in London for multi-site PBR. Two projectors are set up. One shows a video session of the workshop room in Cluj. The other displays a browser on Portia’s computer.
Portia says, “Let’s get started. I want to focus on splitting some items. I’ve invited Zak to join us because he knows quite a lot about this.”
Using a mind-mapping, browser-based graphics tool, Zak starts to create some branches, while discussing with the group.
Afterwards, they use a shared spreadsheet to discuss and write a single example for each of the new split items, so that the people at both sites gain a lightweight but concrete understanding of the details. Later, the group does estimation of the new items, using especially big planning poker cards that can be easily seen by the cameras and video when held up.
Some key points from the multi-site story in LeSS Huge:
Multi-site teams frequently create both obvious and subtle frictions and costs that are surprisingly large in their negative impact.
Qualities that reduce the friction of another site include similar time zone, internal dedicated site (not outsourced), developers that are fluent in the same spoken language, a location and culture that highly values long-term hands-on developer excellence.
A Scrum Master must be co-located with their teams.
Each site must feel like a peer, not a second-class citizen.
Sites must be visited regularly and cross-pollinated.
In meetings, strive for face-to-face with video tools.
The use of shared-document tools make it easy for everyone to modify artifacts together and at the same time.