What’s in your backlog?

Let’s explore what you do and do not put in a backlog. How do these sound?

  • Features and non-functional requirements – Absolutely
  • Bug fixes and change requests – Yes, probably
  • Risk avoidance and risk reduction activities – Sure, maybe
  • Opportunity exploitation activities and marketing ideas – Now you’re just getting weird!
  • Team building and social events – Erm, no!

Yet, if it’s all just stuff for the team to do, then why not put it in the backlog? Maybe because the customer has not asked for it and the product owner has to own and order it, but let’s look further.

If we used a backlog metaphor for prioritizing backlog work items. It may look like this.

Backlog of Backlog Items

I am not suggesting these are the correct elements for including in a backlog, I am just showing the common ones. However, I am probably getting too abstract too quickly. Let’s start at the beginning.

A Backlog Primer

For agile teams, backlogs represent their To-Do list of work. All the things they need to complete before the product or project is done. Now, there may be interim releases. In fact, there should be interim releases delivering valuable functionality as soon as possible. However, there typically remains a list of remaining work. For long-lived products this list may never be emptied by the team, instead it is refreshed and reordered based on the latest priorities.

While the team works from the backlog, it is typically prioritized by a product owner / business representative / ambassador user that sequences the work. This product owner manages the backlog, keeping things up to date with the latest product decisions. They also flesh-out items prior to work starting on them. Product Owners also answers questions about the work from the team, etc.

Here’s a typical backlog showing a combination of features, change requests, bug fixes and a couple of risk reduction activities.

Backlog example

Types not Granularity

This post discusses the types of things in a backlog, not the names we give different levels of granularity. Big chunks of work might be grouped into releases and then divided into themes, or features, epics, user stories and tasks as they get smaller and smaller. There is not an agreed to hierarchy at the large end of the spectrum, often teams miss out one or two of the theme / feature / epic options. However, most teams use the user stories and tasks as work gets smaller.

Nevertheless, this post is about types of work, regardless of their size or what we call them.

Scrum Product Backlogs and Sprint Backlogs

Your view of a backlog may be different from mine. Most people I meet these days were introduced to backlogs through Scrum.

The Scrum Guide describes the product backlog as an ordered list of everything that is known to be needed in the product. It is also the single source of requirements for any changes to be made to the product. The guide goes on to describe the sprint backlog as the set of product backlog items selected for the sprint, plus a plan for delivering the product Increment and realizing the sprint goal.

My backlog history goes like this…

“It’s All Just Work We Have to Do”

I was first exposed to backlogs of work in the early 1990s. Working as a developer at Data Sciences Ltd in the UK I wrote a program to manage our work tasks on a government project. My project manager saw it one day and two interesting things happened.

  1. He did not chastise me for working on a side project of developing a visual work tracker rather than working on the client project.
  2. He asked why it did not contain all our bug fixes and change requests? I did not have a good answer, other than those are different buckets of work we should track separately. He dismissed this explanation and told me to add a flag if I wanted to track work types separately, and said “It’s all just work we have to do” and walked off, but his insight stuck with me. The class of work is secondary – all this stuff needs to get done.

My visual work tracker was quite limited and I abandoned it. The database connection from Easel (a language better suited to building graphical UIs for mainframe systems) did not support concurrent users well. Yet, a couple of years later when we started creating DSDM I knew the backlog was “Just work we have to do”. The backlog is the input-hopper for team work. The product owner is the input-hopper custodian, often subject matter expert, and settler of priority and compromise disputes / negotiations.

 

Risks in the Backlog

I have been keen on proactively addressing risks for many years. Just as features deliver value, risks in the form of threats to the project cost money and cause delays - if they occur. As such, these threats are potentials for anti-value. Like bank deposits and bank-fees, the act of adding value and avoiding losses go hand-in-hand to maximize value.

In the late 1990s I used RUP with some clients and was impressed by the Elaboration phase’s goal of tackling risks early in the project lifecycle. I corresponded with Philippe Kruchten, co-author of RUP, about how to illustrate the good work done on risk reduction during Elaboration that often did not have a lot to demo or show for it. I ended up creating Risk Burn Down graphs for my projects. I wrote about these ideas when I started blogging in 2006 as Risk Profile Graphs. By this time I’d been using them for 4-5 years and knew they were well received by sponsors and executives.

Later in 2006, I wrote about risk adjusted backlogs and Agile Risk Management explaining how to insert risk avoidance and risk reduction activities in the backlog. In 2012 I presented some Collaborative Games for Risk Management at the Agile 2012 Conference and PMI Global Conference. 

When members of the project management community read these posts and papers they correctly criticized my ignorance around proper risk management terminology. Risks, of course, can be positive (opportunities) or negative (threats). I was only talking about negative, potentially harmful risks (threats) when I talked about inserting risks based activities in the backlog. A real risk-adjusted backlog has both threat avoidance and reduction steps, as well as opportunity exploitation and opportunity enhancement actions.

This is how we got to risk avoidance and opportunity exploitation activities in the backlog.  One aims to avoid costs, the other aims to generate new value. Risk management techniques like Expected Monetary Value converts probabilistic events into financial values. For instance, if we have a 50% risk of incurring a $400,000 loss then this event’s expected monetary value is 50% x $400,000 = $200,000.

Likewise, we can assess opportunities too. If the average profit for new customers is $20,000 per annum, then we can determine if inviting qualified applicants on a factory tour with product demos and giveaways that costs us $500 per head is worth it at a 5% conversion rate. Here $20,000 x 0.05 = $1,000, so yes, it appears worth offering factory tours and giveaways for qualified leads.

Multiplying guessed benefits or losses by guessed probabilities is an inexact science. However, it is one that the insurance industry has spent centuries trying to master. So they err in their favor and often price based on what the market will bear. Yet it happens throughout all forms of business and is the basis for taking an economic view of production that underpins all the return on investment and prioritization schemes such as Weighted Shortest Job First. We are constantly looking to maximize value.

So, if a team building lunch is important for boosting performance or reducing the likelihood of conflict and delay, why not put it in the backlog? If it would be helpful for someone to walk the VP of sales through the latest product demo, put it in the backlog.

The Product Owner remains the custodian of the backlog, but with some discussions around threats and opportunities, they often see the advantages of adding these other work types to the backlog. Taking an economic view of work allows us to decide “Where is the next best dollar spent’. That may well be on Feature X or a site visit to help build relationships and increase motivation.

 

 


AI Assistants for Project Managers

Robot hand
Predictions like “AI will take our jobs” sound scary. However, long before our jobs as project managers are taken, AI will help us. In fact, it already is, and we don’t think about it much. While writing this article, AI in Microsoft Word and the add-in Grammarly helped protect you from the bulk of my spelling and grammar mistakes. This is how AI will help us first, by doing small things we are error-prone with, before tackling larger tasks.

Like me, do you spend time booking meetings, finding rooms, and distributing information? Do you analyze backlogs and scope outlines for potential risks, or review estimates for commonly missed activities? Artificial Intelligence (AI) can help with these tasks and many more.

Imagine having a non-judgemental expert monitoring everything you do (and do not do) at work and making helpful suggestions to you in private. This expert is constantly learning, is plugged into all the latest research and works for free. This is the not too distant future of AI assisted project management.

June was Technology month at Project Management.com, and there have been a few articles about AI taking away project management jobs. This article focusses on ways AI can help project managers which will happen as AI develops and before it can replace jobs. It deals with automating the process and science parts of project management, leaving people more time to focus on the relationships, leadership, storytelling, empathy and emotional intelligence side of projects that are harder to tackle and are (currently) best done by people.

AI has come a long way since Microsoft rolled out the annoying and not so helpful “Clippy” Office Assistant tool in 2003. It was never tuned for project managers, but it if were it might have looked something like this:

ClippyInstead, AI is becoming more sophisticated and useful. Gmail will remind you to attach a file if you mention “attach” in the text of an email that has no attachment. Most people use personal assistants like Siri and Cortana on their phones, or Alexa in their homes. Voice recognition and comprehension are steadily increasing. Google recently demonstrated their new Google Assistant calling and interacting with a hair salon to book a haircut. Clearly, these tools will soon be ready for prime time and their use will be widespread.

Kevin Kelly, futurist and founding executive editor of Wired magazine, says in his TED talk: “Everything that we have electrified, we are now going to cognify”. In other words, we will add intelligence to devices and products. Kelly went on to say, “I would suggest that the formula for the next 10,000 start-ups be very, very simple: take X - and add AI.

To understand how AI can help project managers, let's examine its basic capabilities.

  • Knowledge Based Expert System (KBES) – these work from decision trees of IF - THEN statements to provide expertise. Gmail’s attachment reminder works with similar IF body_text includes “attach” AND Attachment = False THEN issue a warning.
  • Artificial Neural Network (ANN) – these systems model our real brains and consist of networks of weighted connections. They can be programmed to learn, recall, generalize and apply fuzzy logic. So, if we teach it someone 4ft high is Short and someone 6ft high is Tall it can generalize that someone 4ft 6 is “Not very tall”. Being able to make these types of generalizations are important for realistic interactions with people, such as Google Assistant making a hair appointment.
  • Machine Learning – this builds on Knowledge Based Expert Systems and Artificial Neural Networks to create predictive analytics that can provide validation and advice. In the project management space, this is the technology that can help with checking for missed risks, rebaselining plans, recalculating the Cost of Delay for waiting initiatives, etc.
  • Chatbots - AI powered programs designed to simulate a conversation with humans. Chatbots use artificial neural networks and machine learning to combine domain intelligence with natural language processing. This gives the impression of interacting with a (currently somewhat) knowledgeable person.

If these technologies sound far-fetched in the project management field, consider the quote “The future is already here — it's just not very evenly distributed”. Agile tool vendor Atlassian, already provide project assistants that help with budgets, estimates, and sprint management. They also have chatbots to share project information and remind team members for estimates and status updates.

Moving forward, these tools will be expanded to help check our work for common mistakes, just as Word checks for common spelling errors. Every industry has catalogs of defect origins and removal methods (here is one for software projects) AI assistants can apply this knowledge and suggest steps to help avoid or reduce these risks. It is not an exact science and as a project manager, I may choose to dismiss potential risks flagged. However, having assistants available to highlight these risks or list the top 10 estimation omissions in my field is probably better than not having them.

AI assistants can also alert project managers to slowly developing trends that might otherwise go unnoticed. The old saying that projects become late one day at a time is very true. Optimistic project managers with “Can-do” attitudes often underestimate the impact of small setbacks and or hope that teams will “catch-up” later.  This hardly ever happens, and AI assistants can be programmed to alert early and avoid hope-based-planning.

Over-Reliance?

There is a risk that with expert knowledge systems, organizations may be tempted to use inexperienced project managers. Or project managers become reliant upon these tools and not think as deeply as they may otherwise. Like any technology, a fool with a tool is still a fool. However, tapping into standard risk lists from your industry, that gets augmented with those from previous projects in your organization is a smart move.

Having calculators has likely reduced our ability to perform long division calculations manually. However, I don’t want to go back to self-calculation just because I fear an over-reliance on technology. Instead, I want to use technology where I can and free up my time and mental capacity for other work.

Higher Value Work

The PMI Talent Triangle is a good model for thinking about all the work a project manager does. It includes: 1) Technical Project Management – the project mechanics described in the PMBOK Guide and Agile frameworks, 2) Strategic and Business Management – your industry-specific work, and 3) Leadership – the people dynamics of projects.

If we squash the triangle out and lay the pieces in order of how much impact the project manager’s contribution has towards project success we get: Technical, then Strategic, and then Leadership. By this sequence, I mean that if the basics of Technical Management are met then Strategic and Business Management work is more significant. Furthermore, good Leadership has an even greater impact on overall project performance than Strategic and Business Management Work, and Technical Project Management.

This sequence is shown below:

AI Focus

The good news for us as project managers is that (currently) AI is best suited for the lower value end of this work spectrum. It is already capable of assisting and saving us time with Technical Project Management work. Next, it should soon be commonplace to get AI assistance with Strategic and Business Management tasks. This will involve accessing machine learning focussed on our industry domains, like ROI models, common risks, and estimation omissions.

The last area AI will move into is the Leadership domain. Machine learning requires deep data sets in a consistent form to draw reliable conclusions. The people dynamics of motivation, conflict management, and negotiation are harder to classify and rank.  Currently, most people would rather work with a real person to solve issues or discover their calling. Who knows, maybe in future people will prefer to interact with chatbots who’s decision parameters can be shown to be neutral and fair. This might be preferable to dealing with people with all their inherent bias and gaps in knowledge.

All I know for now is that I currently welcome any AI assistance I can use. It is likely to safeguard me from making basic technical project management errors or omissions. It should also be helpful soon in providing industry knowledge and best practice – like having a seasoned professional in the industry available to look over your work. However, AI tools will check in real-time before you commit that decision or share a plan.

This leaves me more time to focus on the people. The people sponsoring the project, those working on it, and those who will be impacted by it. They will have their own AI assistants too. Booking meetings, getting rooms, and sharing ideas should become frictionless leaving us to work on the more significant issues.

My recommendation is to stay abreast of AI developments and remain open to trying the tools as they emerge. Standing still in an environment that is moving forward has the effect of moving backwards -which is not good. Where I should probably be more worried is in writing articles like this. It seems like a blend of domain-specific Strategic work with some Leadership based storytelling. Likely a candidate for an AI takeover long before the project manager. (My plan is to get in on the research and get a Chatbot writing this stuff for as long as I can get away with it!)

References:

  1. How AI could Revolutionize Project Management, CIO Magazine, Mary Branscome, January 12 2018
  2. 3 ways AI will change project management for the better, Atlassian Blog, April 7, 2017
  3. Artificial Intelligence in Project Management - Is Your Company Ready for it?, Teodesk Blog, Minja Belic, January 22 2018
  4. AI will Transform Project Management. Are You Ready?, PWC White paper, Marc Lahmann, et Al, 2018
  5. Artificial Intelligence in Project Management, Khaled Hamdy, March 2017

[Note: I wrote this article for ProjectManagement.com, it first appeared here – free membership required.]

 


Agile 2018 Conference – Unraveling Team Dependencies

Agile_SD2018_600x100_Speaking_FM
I am excited to be presenting on the Enterprise Agile track at the Agile 2018 conference in San Diego, August 7. I have worked with several organizations this year that had issues with work dependencies between teams. My session called “Two-Pizza Team Heartburn Relief: Solutions to Team Dependencies” highlights the shift to global rather than local optimization.

We will investigate dependency problems through animations and simulations and then explore some candidate solutions. Each brings their own issues – if these problems were solvable they would have been already, but the suggestions do help considerably. Here is the description from the conference program:

Small teams are great - until they cause bigger problems than they solve. Small teams can communicate more effectively than large teams. They can leverage face-to-face communications more readily and share tacit knowledge without the need for so much written communication. However, for large endeavours, using many small teams present their own problems. Work dependencies between teams can cause major delays through costly hand-offs, mismatched priorities, and blocked tasks.

This workshop introduces strategies for structuring teams to reduce hand-offs and dependencies that create blocked work and delays. By investigating the (lack of) flow through multiple teams we can diagnose the cost of hand-offs and culprits of delays. We examine tools for making hand-offs and dependencies visible to highlight and bring collective attention to the problems. We then explore resolution patterns and work structures that maximize small team communications but limit negative aspects of managing multiple, inter-dependent project teams.

Learning Objectives

  • Understand the time and cost penalties of team dependencies and hand-offs
  • Gain tools for making dependencies, queues, and blocked work visible
  • Learn how and when to balance small team benefits with more dependency issues
  • Share implementation patterns and strategies to maximize team throughput
  • Examine the pros and cons of larger teams, feature teams, and product vs. project development.

That probably sounds more technical than it really is. It is a workshop to show people how teams often get stuck with work items when they rely on work from other groups. It combines anecdotes and experiences from 20+ years of agile consulting along with some insights from Troy Magennis on dependency delays, and Dominica DeGrandis, author of Making Work Visible.

Through case studies and exercises, we explore the hidden impacts of well-intentioned small teams. First, we’ll explore the “mostly harmless” two and three team dependencies, and then see the impacts when five or six dependant teams try to get work done. Please come along if you are attending the conference and have issues with dependencies between teams.


Mining Bitcoins with Your InstantPot: Has Agile Popularity Gone Too Far?

InstantPot_Bitcoin_MachineI am excited to be presenting at next week’s PMI-SAC Professional Development Conference on May 2 in Calgary. I am looking forward to seeing old friends and meeting new people at this new format conference.

My session “Mining Bitcoins with Your InstantPot: Has Agile Popularity Gone Too Far?” examines the hype, realities and use of agile approaches.  Here is the presentation summary from the conference site:

“The project world seems to have gone agile mad. The PMI has stuffed agile into everything in an attempt to stay current, but is it right or even helpful? Fear of missing out has vendors claiming to be agile and executives asking managers to be more agile. However, like InstantPots and cryptocurrencies, does the reality live up to the hype?  This session investigates the agile trend and looks at a new breed of agile suitability filters that help organizations apply agile approaches only when and where they make sense.

The presentation profiles organizations that effectively use agile approaches and how to build PMOs that support agile, hybrid, and traditional project teams. We look at the limitations and applicability concerns of using agile approaches. It would be naïve and arrogant to believe that agile (or any other approach) is universally the best way to execute projects. So, what types of projects do suit the approach (spoiler alert: novel, knowledge-worker projects) and what types of projects should best avoid it?

We will explore hybrid approaches where certain portions or periods of the project can exploit agile approaches. In these hybrid scenarios, we examine how to coordinate and integrate the agile work with the more traditional, plan-driven work. Also, we will examine the cultural fit of agile approaches and investigate how corporate mindsets and values effect structures and decision making. Trying to force fit a new mindset that is at odds with the corporate culture will inevitably be met with resistance. So, we need to be smart about what we try to introduce and how we plan to do it.

Our end goal should be successful projects and happy stakeholders. The approach we take should be appropriate for the tasks at hand, there are no points for style or conformance, since every project is different. So, learn how to analyze the project variables that include organizational, standards, technical and team components then act intelligently within that framework also guided by the inputs of the contributors.”

That description was for a 2-hour slot proposal and I was assigned a 1-hour slot so I will have to cut some material, but I will cover the highlights. The presentation title and outline are deliberately a little controversial to hopefully spur some reaction and interest in the session.

Having been involved in the development of the recent PMI standards, I personally do not believe “PMI has stuffed agile into everything in an attempt to stay current”, however, I can see how this may appear so to external parties.

As agile rides the hype cycle it is important to retain a grasp on where it can add value and where it’s use should be limited. This session aims to strip away some of the hyperbole and ground people with some practical application tools. Please come say hello if you see me at the conference!

PMI-SAC PDC Banner