Four Ways To Improve Predictability (Tip #1)

I’ve spent the majority of my career focusing on:

• Delivering maximum value to customers…
• by successfully planning, executing, and delivering on-time, on-budget, high quality projects and programs…
• and building happy, effective, highly productive teams...
• and optimizing the development process for maximum efficiency

One challenge that I've seen over and over is engineering predictability (or the lack thereof). As someone who is passionate about delivering on commitments, little is more stressful than missing deadlines. Improving predictability is a goal that becomes increasingly challenging as projects get more and more complex.

In my most recent role, I led HPE SimpliVity through a SDLC transformation that reduced an 18-month delivery cadence to a 12 week cycle. The team has now delivered 16 back-to-back quarterly releases, on time, on budget, with high quality. The organization has also improved it’s commit to deliver (A.K.A. say/do ratio) from 40% to 120%. Finally, our employee engagement index has increased by nearly 20%.

I don’t pretend to have all the answers; if you're looking for silver bullets, look elsewhere. However, I do have significant experience and insights that I other may find helpful. Thus, in the coming weeks I’ll be sharing a multi-part series focused on the topic.

• You are here --> Post #1: Narrow your planning horizon
• Post #2: Use empirical data
• Post #3: Create smaller teams
• Post #4: Eliminate the fear of failure

It's my hope that this content generates healthy discussion and knowledge sharing!

Tip #1: Narrow your planning horizon

How far in advance does your organization make commitments? Do you have project milestones that are many months in the future? In my career, I’ve seen planning horizons as long as 18-24 months. During projects that long, without fail, those milestones get revised over and over. After leading dozens of projects over my 20 year engineering career, I would never trust an 18 month project schedule labeled as “high confidence”. In fact, I’m convinced that it’s impossible to create a 6-month project plan that will be 100% accurate: delivered on time, on budget, meeting all the original requirements, and with high quality.

Source: Wikipedia

Be certain about uncertainty

A concept called the “Cone of Uncertainty” is my favorite way to illustrate uncertainty in planning. The concept is simple: short term plans have fewer variables/unknowns and thus, work estimates are more likely to be accurate. The further out you plan, the more variables come into play, the cone of uncertainty expands, and thus, the less accurate your work estimates will be.

Source: Construx

Let’s use a cross-country road trip as a simple analogy. For an adventure like this, one likely would build a plan of what route to take, as well as where to stop to eat, sleep, fuel up. Now consider the variability there may be in said plan. For the whole trip there could be any number of unforeseen complications. Maybe there's a detour for road construction, or for unexpected rush-hour traffic. Maybe you get unlucky and get a flat tire. Or perhaps along the way you see some attractions that you want to see, but were not in the original plan. It is highly unlikely that the original plan plays out exactly as envisioned.

On the other hand, it’s far more likely that the first few hours of the trip go exactly as planned. You’re more likely to have up to date traffic information for the area close to your home. You’re less likely to get a flat tire in a short period of time. And you’re familiar with the attractions in your area, so any stops would be well planned.

As the trip progresses and you make short-term adjustments, you will need to adjust your long-term plan accordingly.

Benefits

Possible benefits of narrowing your planning horizon include:

• Increased transparency into project progress
• Early detection of delays, increasing the odds of successful mitigation
• Improved ability to react to unforeseen circumstances, such as changing requirements or project roadblocks
• Fewer end-game surprises (i.e. avoid those projects that are at 90% completion for months on end)
• More predictable short and long term schedules

How to get started

Even if you're not going to fully complete projects in a matter of weeks, you can still benefit from forecasting within a short horizon. Some things to try

• Pick a planning horizon, measured in weeks
• Carve your projects into smaller, independent deliverables (if even for internal, non-production consumption only)
• Replace lengthy one-time planning sessions with shorter, recurring planning
• Don’t over-rotate. The business still needs to know what to expect long term. Regularly refresh your long-term plans, each time you refresh your short-term plans
• Strive for a project backlog that has fine-detailed short term plans, and more coarse long-term plans (see below visualization)

Source: InformIT

Let’s discuss

Do you have any useful strategies for short and long term planning?