Subscribe via RSS Feed

Category: Miscellanous

Stop Padding Your Estimates! There’s A Better Way…

When you pad an estimate, you typically increase your estimated effort/cost due to fear. I call this estimating from a place of weakness. You are afraid of the unexpected event, so hopefully padding it will cover it.

When asked why you think something might take 50% longer or cost more than expectations, it leads to an awkward uncomfortable conversation. And almost always, it creates distrust. How do I know you haven’t padded everything?

A Better Way: Contingency

Instead of padding out of fear or worry, add contingency instead. Let me clarify what I mean.

An example of padding an estimate: Your estimate is for 2 weeks to finish a straight forward project. Because you know stuff happens, you decide to say it will take 3 weeks. That gives you a full week to recover from delays or fix any defects.

How do you know a week is enough padding? You don’t. But you believe you can get away with it easily enough. Saying 4 weeks would be pushing your luck.

Instead, here’s the same example using contingency: You provide your estimate as being 2 weeks. You also explain that typically, there are delays of source material, and defects that need to be addressed before we are all done. So you are adding one week of contingency time, just in case.

You are sharing your estimate and “contingency” from a place of strength and confidence. You show it’s in everyone’s best interest to include the contingency. It builds confidence and trust.

It’s not uncommon for one of my full project plans to have contingency tasks added throughout the plan. If the customer wants to push back, here’s how the conversation goes:

Customer: “I see every time you are working with the API team, you show a task for 3 days, but you add a contingency task of 3 days as well. Can you explain what that’s for?”

Our team: “Typically, 3 days is enough time to integrate the APIs and move on to the next task. However, in the past, this API team is always overwhelmed with work, and some of the APIs are new and complex. So the contingency time is based on history and allows us to plan for those delays. If they can actually reduce turn around time and hit the 3 day mark, we’ll be ahead of plan and deliver early. After all, there are 5 API tasks and 5 contingency tasks.”

Customer: “I’m not sure I am comfortable with this approach. I think we should remove the contingency tasks.”

Our team: “We can do that. However, if in fact those teams don’t hit their tasks, we are going to say ‘We Told You So’ and you can’t hold us to the dates!”

And no we don’t say it this way – but that’s the message.

At this point, I can’t remember a time when the customer didn’t agree to keep the contingency in the plan. After all, they want to be successful too.

In summary…

Don’t pad your estimates. Clarify why you are padding by adding contingency and sharing the risk. Hope this helps you with your estimates. Would love to hear your experiences as well.

Do you have an opinion on why some software projects fail to deliver?

Survey Says!

If you have an opinion on why some software projects fail to deliver, then you’re invited to participate in our online survey.

Take just 5 minutes  and we’ll send you a $5 Amazon gift card (details on the survey site)!

Start now: http://www.surveymonkey.com/s/geneca-mytwocents