How to estimate time required and deadlines for creative and problem-solving tasks?

Question

How to quantify time requirements for tasks that are almost purely problem-solving or creative? And how to set deadlines based on those estimates?

Practical scenario in my team:

Steve must find a solution to a never seen before problem A, which includes researching/reading up about similar problems, analyzing the problem A and trying out various ideas until he finds the solution. The company wants problem A to be figured out and solved by end of this week. At the same time, there is another task B which we know will take around 30 hours, and that needs to be done by end of next week. Senior management wants regular updates on status of B and is not keen on doing it last minute.

Both A and B must be done by Steve.

How exactly can I quantify A so that I can, for example, ask senior management that we start B later?

Gut feeling says researching and figuring out what A is may take a whole day, implementing it four-five days. Could I then say that A will take six days, after which we can dedicate four full days to problem B?

Any other ideas how I can tackle a problem like this?

Answer 1

Many times you JUST CANNOT predict how long some things will take. Even the most experienced managers in organizations with infinite amounts of money at their disposal frequently FAIL to meet deadlines when attempting something new. This is especially true when you're dealing with small groups who have a lot of responsibilities on their backs to begin with.

Steve needs to triage both problems and report back on the feasibility of doing each one or at least indicate what portion of each one can be realistically completed in a estimable amount of time.

Your job is to back-up Steve and make upper management understand that project A is an unknown which requires some discovery before an implementation can be planned.

Due dates are rarely carved in stone. It is up to you to make sure that your people are diligent in meeting the challenges. You also have to set expectations with your bosses-- this means saying "no" to unrealistic demands.

Some managers think that being able to predict completion dates is their prime priority. That's only half (or less of) the job. The other half is removing obstacles. What are you doing to keep your team's velocity up? What obstacles are you removing? What resources are you providing?

Answer 2

Your question is how to quantify your estimate to allow management to let Task B start later.

My advice would be to set the expectations.

You know each Task is going to take around a week. This leaves no contingency and assumes he has no distractions or illness.

You could use a Three-point estimation technique to give a Best Case, Worst Case and most likely.

After this let your business know you will provide an update refining the Task A estimate on the 3rd or 4th day after an update from Steve.

Answer 3

TL;DR

• If everything is "top priority" then stakeholders have failed to prioritize.
• Management targets are not guarantees that a goal is achievable.
• Estimation by proxy rarely works; estimates should be provided by the person who will actually perform the task.
• The is no magic formula for delivering undefined work packages on time.

Senior management is responsible for clearly defining priorities. The project manager is responsible for tracking the project and providing senior management with information about the status of, and risks to, the project. Poor Steve (the person mentioned as the sole resource in your post) is responsible for providing you with reasonable estimates of how long it will take him to complete a clearly-defined task, and then slogging through one task at a time in whatever order senior management decrees.

Use a Story Spike

The company wants problem A to be figured out and solved by end of this week. At the same time, there is another task B which we know will take around 30 hours, and that needs to be done by end of next week. Senior management wants regular updates on status of B and is not keen on doing it last minute.

When you have a dependency of unknown size, you can use a story spike. A story spike can be time-boxed, and the results of the spike can be fed back into your project plan and your estimating process.

Your Next Steps

In your specific example, you have a number of management targets with fixed delivery dates. However, you also have scheduling dependencies on a task that has not been properly estimated. To address these issues, you might take the following steps:

1. Inform the stakeholders that Task A can't be estimated yet.

   Whether getting it done by the end of the week is a management target, not an indicator of whether or not the goal is reasonable, and this difference requires two-way communication and some education of your stakeholders. Estimating Task A is a missing milestone in your project plan, and you need to make time in your plan to accommodate the activity.

2. Inform the stakeholders that Task B can't be started until Task A is completed unless the stakeholders want to re-prioritize.

   Assuming that there are no dependencies between tasks A and B, if Task B is the most urgent then it should be done first. In such cases, revise the project schedule and perform Task B first. On the other hand, if there is a dependency or if Task A really does take precedence, then stakeholders simply must accept that Task A must finish before Task B can start.

3. Allot some percentage of the current week to your story spike.

   If you have four days left in the week, and a fixed delivery target of Friday afternoon for the feature, then a reasonable approach would be to allocate one or two days to researching Task A in order to estimate whether the task can be completed within the target time frame.

4. Perform Task A or escalate.

   If the spike indicates that Task A is feasible within the desired time frame, spend the remainder of the week performing it. Otherwise, provide as accurate an estimate as possible to the stakeholders, along with identifiable schedule risks to other tasks such as Task B, and let the stakeholders figure out which task is more important to complete first—or whether Task A should even remain in scope, with or without modifications.

Answer 4

I'd get Steve to run A and B is parallel on the first day - a few hours of B, and A for the rest of the day. I want Steve to give me a time estimate for both projects by the end of the day at the latest,

1. Can Steve tell you that he expects that both A and B can be completed by the end of next week?

2. If he doesn't expect to meet the deadlines, can Steve say what kind of support does he needs to complete both A and B by the deadlines?

3. If the resources that Steve needs are not available to help him complete A and B on time, which deadline has more give, the deadline for A or the deadline for B?

Have Steve email you twice a day on his progress completing both tasks and letting you know whether he is still on track about meeting the deadlines. In particular, look for any evidence that Steve is stalled and needs some backup.

Follow-up comment from @teego1967 "One sign that a project is going to hell in a handbasket is a requirement for "realtime" email updates to management about progress. Just briefly chat with Steve to make sure he's on track, that will do the job without wasting his valuable time. Requiring frequent status-update emails is a major demoralization. It signals a lack of trust."

My boss required regular email updates from me on my projects - nothing verbose - just to make sure that I wasn't getting mired somewhere without assistance. I am not aware that I ever got demoralized from this requirement. In turn, my firm's clients also required my boss to give them regular email updates and my boss complied. No skin off his nose either. It takes three minutes to put a status email together and send it. What do you mean. "wasting Steve's valuable time?" I take more time to wipe my nose and stare out the window. Reagan used to say: "Trust and verify" - Following up on action items is a basic, fundamental tenet of supervisory management. Trust or lack of it are irrelevant to follow up and verification, but due diligence have plenty to do with both. No manager that I ever worked with likes nasty surprises. And I have done enough management to hate it when a nasty surprise happens to me.

Answer 5

"How exactly can I quantify A...?"

I would split A in 2 stages - research and implementation, and would give the management an estimated deadline for the research stage. One of the results from the research should be a plan how to solve the problem, with estimated time needed.

Regarding the order of work - it depends on the priorities of the company about these particular tasks - are there customers waiting for the results of A or B, are there any relations between A and B, the overall impact of either task, etc.

If it is imperative that both tasks are completed by the end of the week, I would look for a way to help Steve. For example, if Steve is the only person with specific knowledge required for A and B, are there any subtasks that can be done by someone else - like browsing internet and downloading/bookmarking potentially useful information, so Steve can concentrate on the research. Or preprocessing some data, or making the interface nice.