Agile Product Strategy, Marketing, ProductCamp and more.
RT @celia: (via @StephenFleming) Steve Blank: "The golden age of Silicon valley is over..." m.theatlantic.com/business/archi… @sgblank
Scrum
Should the product manager be product owner for new product development in Scrum?
2243 days
In an earlier post, I wrote that the model of having a strategic product manager partner with a tactical scrum-team-immersed product owner delivers business value best under limited conditions. As this will be one of the most controversial topics during my talk tomorrow on #ProdMgmtTalk, I want to reiterate my rationale here in a more verbose form.
The two conditions I listed are:
1) The product is marketed to a segment of customers, not just one.
2) The product is already released to market, and is thus beyond v1.0.
The "Product Owner" without Product Management responsibility is really a Backlog Manager.
Delivering business value in this context means developing a product that meets the needs of the market, and fulfills the product vision that capitalizes on market opportunity. In the stage-gate days, product managers would “capture” that vision in MRDs and PRDs and toss them over the wall, and then be frustrated that the product which came out the other end did not deliver. By providing periodic visibility of work product, Agile processes are often implemented to address this long-observed waterfall limitation.
I suggested that in other conditions–specifically mass market products during intitial development–companies may wish to consider having the product manager serve as product owner, in order that the person with the best knowledge of the market and problem space is guiding the product’s development. This blog is an attempt to clarify some of the arguments supporting this position.
My suggestion generates much contention among ISV product management purists. In agile development in general–and scrum, specifically–heavy demands are placed on the product owner. Chief among these demands is the regular management of the backlog by breaking down high priority user stories, and the need for consistent availability by the development team. Product owner availability ensures the development team has the resource necessary to explain the rationale for features, and to work with team to validate the best way to implement each user story.
Product managers are members of the business, not of the technology team. They are responsible for identifying market problems that people will pay to solve, which can most effectively be done by spending time with customer representatives and studying the market rather than by spending time in the scrum room. Product managers must interface with the rest of the business in order to coordinate the surrounding roles: creating training guides, helping marketing with positioning and messaging, sales support, etc. Product managers are often responsible for more than one product, leading their focus to be strategic rather than tactical.
Clearly, these two roles cannot be the same person! Or can they? I believe there’s a case for these two roles being occupied by the same person during the initial release of a product.
When companies implement scrum, common guidance states that product management needs to staff up by adding a product owner to the ranks to support the product manager. I believe in some company cultures, this remains the best option. I believe, however, that an alternative model should be considered: While release 1 of a brand new product is under construction, the product manager works only on that product and serves as product owner on the scrum team personally.
This proposal implies a redistribution of responsibility among the product management team — someone else must manage the product manager’s other products to allow the product manager singular focus on the new development. Many teams won’t be able or willing to make this sacrifice, for one or more of many reasons including (1) there aren’t market experts to replace the product manager on the other products, or (2) the cost of bringing in a product owner on the new product is lower. However, there is a primary disadvantage to keeping the product manager on multiple products including the release-1 effort: Lack of Focus.
A Product Owner is often brought in to partner with product management as part of the “product owner team” or “product discovery team“. This arrangement usually involves the Product Owner meeting regularly to ensure day-to-day execution of the product manager’s vision, while the product manager meets with customers, develops pricing, champions the product, and manages the other products in the product manager’s portfolio. Conversely, the Product Owner may not have the luxury of all the customer visits, market research and competitive research the Product Manager used to develop the business plan and the product vision.
This lack of focus is the problem.
The Product Owner is encouraged to be in the scrum room during a majority of his time, to help with all the day-to-day decisions that go into building the right product. In this arrangement, this means that the person with the deepest direct knowledge of the market and the problems the product is intended to solve…isn’t. When called upon to explain rationale and help with interaction design, or any of the other elements about which the product owner is asked to participate, the product owner must rely upon limited customer interaction and second-hand understanding obtained through others. This leads to weaker positions in discussions with developers, and may have unintended consequences that manifest in the design of the app. These consequences may not be related to something the product owner says–in fact, it may be because of something the product owner doesn’t say.
To implement this, it may require a bit more sequential thinking than scrum purists would support:
- The product manager builds out market justification and designs the solution (with the lead designer) before starting construction.
- The product manager allocates a majority of time during construction to development team support.
- During buildout, the product manager strategically allocates mid-sprint time to customer visits and product reviews. This time after working through the current sprint user stories and before the race to capture points at the end, is often a quiet time for the product owner. Note in a two week sprint, this may be roughly 4-5 days of a 10 day sprint.
This last point illustrates why this requires allocation of the product manager to the new product exclusively — there isn’t leftover time in this model to work on other products. When exclusivity is not an option, the product manager works with the product owner to ensure the solution addresses the market problems. As I’ve mentioned, this separation is risky.
One way to counter the risk is to bring the product manager in when anything important comes up, however, it’s very difficult to determine which instance of the “butterfly effect” may require the presence of the product manager. Often the product manager is unavailable, and delaying an answer to arrange a conversation may interfere with the team’s velocity.
Another option is to have the product owner conduct the competitive research and customer interviews to ensure they are market experts as well. But then, isn’t that blurring the lines and having the product owner serve as product manager? And if they’re showing prototypes of the product to customers throughout the construction process, are they as available as the development team needs them to be? Perhaps teaming with a product marketing manager can offload some of the outbound marketing and sales responsibilities, but there will still need to be someone with product and market knowledge guiding the sales of the product and its positioning in the market. And the product owner in this case will be doing a lot of the inbound marketing that is the core of product management.
Are there other approaches? Can these challenges be resolved by having another member of the design team–the user experience specialist–available to the team as well?
Please share your thoughts in the comments.
Agile Product Management and the Product Owner Role – A Town Hall
0265 days
Many of you know me as a frequent participant in the product management community on Twitter, and as an organizer of ProductCamps. After participating on the organizing committee of four ProductCamp unconferences in Austin and two in Atlanta, I’ve just submitted my first session proposal for a ProductCamp — this one for ProductCamp Atlanta. I’m excited to contribute more, in a different way, to the product management and ProductCamp communities.
My topic is “Agile Product Management and the Product Owner Role – A Town Hall.” This talk is in response to the ongoing confusion that exists as engineering teams have moved to agile frameworks, while product management struggles to figure out how it can fit in. In my view, the benefits to the business of agile frameworks are equal to if not greater than those delivered to engineering, so it is in product management’s interest to optimize this evolution.
Some of my views can be found in earlier posts. I’ve touched on the confusion on several occasions in the past:
- http://johnpeltier.com/blog/2011/07/22/where-product-owner-as-backlog-manager-fits-best/
- http://johnpeltier.com/blog/2011/03/22/prodmgmttalk-product-manager-vs-product-owner/
- http://johnpeltier.com/blog/2011/02/25/thoughts-on-agile-product-management/
In my session, I’ll summarize common organizational models found in companies today, and lead a discussion bringing in the experiences of the group. From this exercise, I hope that we’ll all come away with new insights into the way product management can deliver incremental value when partnering with an agile engineering organization.
Where “Product Owner” as “Backlog Manager” fits best
9302 days
In a recent post, Saeed Khan wrote that the Scrum role of “Product Owner” should be redefined to describe a more accurate set of responsibilities: “Backlog Manager.”
Saeed’s argument is that the Product Owner role is really a project-specific role implemented within the Agile movement as a solution to inaccessible product management. Saeed observes that more and more product management responsibility is bleeding into the Product Owner role, which is intended to be the funnel of business input to the development team. This conflicts with the primary responsibility of the role, which is to provide a set of fully groomed user stories representing the next most important features that should be built, and to be available throughout the sprint to answer questions.
After considering this since Saeed’s post, and based on my own experiences on four scrum projects, I’ve come here to propose that there is a “best case” condition under which this arrangement makes sense, and several others under which it doesn’t quite work.
The Case where a Backlog Manager makes sense
The first thing to remember is that Scrum has infiltrated software development from the custom software world. When building custom software, there is a single customer whose wants and needs are primary — and time has shown that having a representative of the customer in the war room at all times can help the development team deliver something meeting the customer’s needs in an efficient manner. In this context, because the Product Owner is often an employee of the client who has broad and wide domain knowledge, it makes sense that a single person bears all the responsibility of the product–and it makes sense that a single person can deliver it! This set of conditions does not demand that we separate the tactical “product owner” responsibilities from the business “product management” responsibilities because they are so intricately tied together–and the product is being built for one customer!
When we look at independent software vendors, however, we encounter the difficulties with balancing the strategic and tactical that many Product Managers express. A product manager must synthesize the needs of a market segment to deliver a compelling offering, and successfully prepare it to be marketed to that segment. If the Product Manager (title) occupies the role of Product Owner, spending much of his/her time on tactical items for upcoming sprints, the pattern seems to be that access to the market is less common.
It is at this point that the notion of “Time” becomes important. I believe the tactial responsibilities of backlog management can be offloaded by a Product Manager to someone filling the role of Product Owner (a Technical Product Manager, a Business Analyst, etc) under one condition: After a product’s initial release to the market.
In my ideal process, before starting construction of a new product, the Product Manager will spend significant time in the market understanding the problem space, developing business case and initial requirements, to the degree that he/she is prepared to spend a number of months focusing on development. It is only with this focus that a clear vision of the product can be brought to the team and shepherded to completion.
Backlog management will take a great deal of time, but the product manager’s command of the entire product is essential to having it built correctly. With a competent team, the Product Manager serving as Product Owner can still find time to make periodic visits to customers to get feedback on early builds and gain the input needed to “inspect and adapt.” But if this feedback only reaches the team second hand, the risk to the cohesiveness of the product is too high.
After version 1.0 of a new product, when the product is in customers’ hands and is being considered by prospects, the vision, business case and purpose of the product is already established. Now that the product is built, the product manager is looking for additional market segments and business opportunities, finding the problems which adding solutions into the product would open up new markets. Additional important and complex features may be added, but the basic nature of the product is already set. Given this stability, the profitability of this product can be managed by one individual, while the tactical implementation of new features can be efficiently offloaded to another. This is the case where separating responsibility between a Product Manager and a Product Owner makes sense.
These three statements summarize my position:
- For custom software vendors, there really is a single “product owner” that fits with Scrum orthodoxy. This “Product Owner” is a representative of the customer.
- For ISVs developing version 1 of an off the shelf product, a single person should own the business and technical sides of the product. That single person learns the market and helps build the right solution.
- For ISVs developing later releases of an off the shelf product, the backlog management responsibilities can be split from P&L / business responsibilities. The boundaries are set, and a “backlog manager” or “technical product manager” can work with the Product Manager to understand the goals of the next release and execute, while the product manager coordinates sales, channels, support, and other arms of the company.
Special Note: During the development of a Version 1.0 release, because it makes sense to have the single person with best understanding of the business case overseeing the development effort, in most cases it does NOT make sense to have one person overseeing the buildout of more than one product! The workload to understand business needs, positioning, usability needs, requirements and user stories, and implementation details is too great to split that person’s time into multiple threads. I can confirm this from personal experience!
ProdMgmtTalk: Product Manager vs. Product Owner
8424 days
Today’s Global Product Management Talk, a weekly realtime Twitter conversation among product management professionals, covered the topic “To Agile or Waterfall…Does it Matter?” featured John Mansour, CEO of Agile Bench. As an agile product manager, I looked forward to this conversation despite that its timing at 6pm Eastern prevented me from participating realtime. The first posted question was “Product Manager, Agile Product Owner, What’s the Difference?” Since this is near and dear to my heart, I’ve composed my response here.
Some of the comments from the host that I found especially helpful included:
- agilebench: “Product Manager and Product Owner are two roles, sometimes occupied by the same person.”
- agilebench – PMs – outward (customer) facing. Channel, brand, price, the whole product. Strategic. POs – inward (project) facing. delivery, detail focused. Tactical.
The Problem: One huge problem is when one person tries to do both roles. From personal experience, I can confirm that especially with large-scope products in multi-national corporations with products in multiple silos, it’s practically and effectively impossible. Brainmates summarized the problem I experienced: “Market focus suffers, because it’s easier to work on tactical development activities.” In companies I’ve been exposed to, it was not only “easier,” but there were so many fires that there was really no choice. Because of the product’s complexity and the complexity of the organization (not to mention infighting and territorial-ism) there simply wasn’t enough time to participate in user story level design discussions, groom the backlog, coordinate sales and support, and simultaneously monitor the competition, research trends and pricing, and the other activities a business leader in the company should perform.
The Team Approach: As an alternative, agilebench suggested “can you find someone with product sympathies who can act as a proxy for you?” This echoes an approach I’ve seen becoming more commonly cited–including last week’s Technology Association of Georgia presentation by Mike Cottmeyer–of having a Product Owner Team. The team consists of a handful of roles that must be addressed: Product Manager handling the outward market-facing strategic functions, product owner acting as the development liason and sitting with the development team and driving home the detail around the user stories, and potentially a user experience analyst and others. Thus as Roger Cauvin proposed, “one model is that product management is more market facing, while product owner is more inward facing.”
NOTE: Another great read about a similar topic was recently posted to On Product Mangement.
The Balancing Act: One major concern about this approach is “too many cooks in the kitchen,” or not having a single point of authority on the product. I think the split approach can work, if and only if the members are diligent about remaining aligned on a daily basis. Another alternative is a balancing act: The PM/PO must be accessible to the team at all times, yet must get in front of customers and the other teams within the company to ensure a successful rollout. This balancing act can work if the team is talented and can work (at times!) without direct input from the product owner. A few calls and meetings spread through the week is one thing, but if the product owner is rarely visible except for the daily scrum, even a seasoned development team may be hard pressed to succeed. Conversely, if the product manager never leaves the scrum room to speak with customers, the team is at serious risk of building the wrong product “successfully”!
Two Pounds In a One Pound Bag: Given more than a full-time job’s worth of responsibilities outside the scrum room, and nearly another within, it should be inherently obvious that a product manager should not be placed in the impossible scenario of performing product manager and product owner responsibilities on two products at once! I can also confirm this from personal experience.
Other Topics: Other topics were discussed, including
- how to gain the respect of the development team
- how to break up deliverables when more than one product manager is on a single product
- when are agile methods not suitable for product management
- How do product managers ensure the vision when agile delivery teams are focused on small pieces of work?
- If it all needs to be done why does it need to be prioritized?
- How do you roll up individual projects into an org-wide roadmap aligned with company strategy?
Plenty of great discussion on all points is archived and available in PDF (start at the bottom!). Enjoy your read!
Book Review: Agile Excellence for Product Managers
0483 days
Agile Excellence for Product Managers by Greg Cohen is, in my view, the best introductory “how to and why” guide to Agile and Scrum for product managers available. There are other great resources touching on specific elements of methodology–such as Mike Cohn’s User Stories Applied–but Agile Excellence puts those techniques into context and explains the benefits of iterative development, the impact of iterative development on how the product manager conducts his activities and why it helps the product manager conduct those activities better than under a traditional development methodology.
Agile Excellence introduces the product manager to agile in a deliberate sequence. Starting with the benefits of iterative development, Cohen covers the specifics of Scrum and then gets into the details of release planning. In my experience, some teams try to “sprint through the backlog” without considering the contents of a specific release — the release planning chapter illustrates a technique for thinking this through to provide marketable value to customers over a series of iterations. To bring the product manager up to speed on writing requirements in the form of stories, Cohen goes into detail about user story management including the types of stories that exist and techniques for splitting the large ones. Examples are provided to illustrate when a story is too big to estimate granularly. This section is not nearly as comprehensive as Mike Cohn’s book, but it’s good enough to get started without lacking the basics.
Cohen shows how to move from ideation to a prioritized backlog, and how the team is able to deliver value to the business early and often by doing so. Agile Excellence walks through creation of a product strategy, followed by a release plan, creation of the product backlog, and finally estimation of the stories using the Planning Poker technique. With a team working on an estimated product backlog, Cohen considers how a product manager interacts with the team and with the business at large (including customers), and discusses signs of trouble and how to react. Finally, for context, Cohen spends a bit of time on XP and Lean development including Kanban.
One of the important points Cohen makes is that Agile and Scrum are essentially a project management approach rather than a development methodology. Development teams need to adopt new techniques to meet the tight deadlines agile demands, but these techniques are not elements of agile. Agile simply assumes the team is “self-organizing” and will make the changes necessary to succeed. The fact that those changes often include pair programming and test-driven-development is left to the development team to discover, however it is useful for a product manager to be aware of them in the event they are evangelizing agile to an organization.
I highly recommend this title for product managers new to agile, and to those needing a refresher. This book puts agile techniques in context, gives enough “in the weeds” level detail, and explains why these techniques are useful and why they make a difference. It’s the book I wish I’d read before beginning to work with a scrum team.
The Product Owner Vision
0503 days
A recent article by David Alfaro summarizes well the core competency of a product owner in scrum / agile development organizations — the ability to consume all the inputs, mix them together and design a cohesive, marketable product that will meet the needs expressed in the inputs. Not surprisingly, this sounds a lot like product management. Here’s how Alfaro expresses this core skill:
After you have listened to all the aforementioned stakeholders the following two hardest steps are ahead: 1) Synthesize all that input into something really remarkable, feasible to implement and affordable for customers. 2) Communicate that synthesis to stakeholders and team in order to go for it.
There are (at least) two major views of this. Design-driven organizations would likely agree with Alfaro, perhaps with the added input and agreement of a design team (artists and interaction designers) on the front side to craft the behavior of the desired product. Others that take a more “agile” approach may guide the product owner to focus on the problems, presenting them to the team in the form of stories and allowing the team to design the solution within a pre-determined set of parameters.
Either way, the product owner’s job is to define with more or less granularity the “what,” and to guide the team to analyze and implement the “how.”
Daily scrum?
0So I find myself a Product Owner running 3 scrums per day, and I have several vastly different teams (except that all are understaffed). One in particular is a high-performing team in which the developer produces great work, but in which there is some frustration with the practice of meeting daily. So I’m thinking strongly about the idea of moving to twice-weekly scrums. There isn’t enough change between days to warrant a dedicated meeting, and the team is vocal enough to raise blockers when they exist.
In my readings about the practice, many are adherents to every nuance of Agile and Scrum, but I have to question this when I am not convinced of the value obtained from the practice. So, at this point my instinct is to try twice-a-week during the next sprint, and reevaluate during the retrospective.
Thoughts?
Recent Comments