How to Make Accurate Time Estimates (Blog)

[Steve Pavlina]

Use this thread to discuss the following entry from Steve Pavlina's blog:

How to Make Accurate Time Estimates

[carenkh]

Thanks, Steve! What you suggest is something I've done in my head for many years, without being as concrete in my thoughts. i.e., I just say "I better give myself more time for that", add whatever 'feels' right (1/2 an hour, 15 minutes), then go with that.

It was quite a while ago that I realized that in my head, I was thinking some tasks would take zero time. For example, let's say I was out with friends, and we were going to meet again later at an alternate location. If I had to stop by my house on the way, I wouldn't add the 15 minutes that would realistically take. It's just stopping by my house! Like I thought I would teleport myself in, grab whatever I needed, then teleport back out. Zero time. Nope! I have to physically park my car, unlock my house, look for whatever I needed, get it, go back to the car, etc. Not to mention the half hour (at least!) you should automatically add to any task if you have small children.

I'm not quite so into time management - I'm a go with the flow kinda gal - I changed my ways because at some point I recognized I was always late to things. It took me a while to realize how rude that is. For the most part, I've been on time for things for years now. Yay!

[overstep]

While preparing a To Do list I use additional colors: green for the time and red for the priority. Sometimes I also draw green and red bars instead of just writing the numbers. It's good to actually see it, that way the truth is clearer. If a task takes three chunks then I draw three bars in one line and mark them separately when they are done.
I also did a permanent list of tasks that I do almost daily, I did it on a spreadsheet, sorted by priority and added a graph. The funny thing is that the most important tasks take the least time for me, they are planning an visualizing. I often skip long tasks if they are not vital along with the saying "There is never enough time to do all very important things, but there is always enough time to do the most important things". There are also tasks that seem long and not vital but they are definitely good an have effect on everything else, one example is a walk in the park.

[Centaur]

I often say in joke, “For a realistic estimate, multiply by square root of two; for a pessimistic estimate, multiply by pi.” That is, I assume a fudge factor of 1.4142 to 3.1416.

[supertom]

shoot, didnt know it took him 3hrs to write a article maybe add a bit more for getting ideas.

[wolverine]

From the student perspective, how would people suggest estimating time for things you've never done before? Obviously, as a student, I learn stuff. Duh. And usually, that stuff is stuff I've never done/known before, or at least not quite the same as previously. This tends to lead, unfortunately, to having to estimate times for projects, assignments etc for things I have no clue about. For example, I'm taking my first programming class this semester, and while I now have a vague idea of how long it will take me to write my assignment codes based on the lab class coding I've done, I still don't really know. And I tend to underestimate, forgetting details and mistakes and stuff, so it can get a bit tight at times (ie, now) when I underestimate for a LOT of stuff all at the same time.

Suggestions, ideas, comments?


Wolverine

[Roman Budzowski]

I struggle with estimating research & development. Those are complex tasks without clear boundaries and estimating then is very difficult, yet you want to know how long it will take.

cheers
Roman

[Jake Birkett]

Roman: And prototyping - I mean that takes as long as it takes to get a decent prototype. Also testing can easily balloon out if bugs are discovered, or the testers get confused by the interface/gameplay and major changes are required. Then there's the polishing phase which you simply have to put a cap on (or set a release date) to avoid it carrying on for ever...

[truthcurve]

I use a combination of two things to estimate my time accurately: how long something will take that I have control over PLUS how long something will take that I have no control over. The latter sounds counterintuitive but I explain it here.

I calculate how long something will take and then multiply by at least 2 depending on what it is. I add to this the amount of time that will be lost to "dark matter," i.e. stuff that is uncontrollable. In both cases, I use rules of thumb. I have never tried to create a fudge factor like Steve suggests. It sounds interesting so I may try it to see how well my rules of thumb have been doing.

Julie

[ThatGuy]

Great post Steve

[Alora]

Last line of the article:
"And remember, this is for posterity, so please… be honest. "

Cracked me up so much I had to join the forums. I also can't believe I just read an entire article about math. Steve is made of magic, if you ask me.

[Keith]

Joel Spolsky posted an article a while back on what he calls Evidence Based Scheduling - essentially a more sophisticated version of Steve's "Fudge Factor". It's built into his Project Management Software, FogBugz.

Evidence Based Scheduling also has a feature I've never seen before: Instead of giving you a specific due date based on the current schedule, it uses statistical analysis to give you a probability distribution curve of when it will be done. So it gives you not only an estimate of when you'll be done, but also the margin of error for that estimate.

Pretty nifty stuff...

[Marth]

Yes helpful article I'll have to read the rest tomorrow.

I find figuring a 5-15 min buffer time helpful to separate tasks when estimating, just to help you not overpressure yourself. You could use about 60% of your 'fudge estimation' instead for the buffer.

[PerDev]

Very Practicle, Great Post

[jabberw0cky]

Very interesting article, Steve. For years I've always multiplied my project estimates by roughly a factor of three. I had never really thought about it much, I had just learned that I'm overly optimistic about what I'm able to accomplish in a given period of time. I think it is a good idea to figure out what my 'fudge factor' is for given activities, to make my estimates more accurate.

For the last several months I've been working with a good friend of mine on some web based time tracking software. They recently released a free personal version of it which I've been using to track my sleep habits and time spent exercising, etc. It's very illuminating to pull a report and see patterns emerge - like the fact that my body requires about 6.5 hours of sleep regardless of what time I go to bed. I'm planning to now use it to help track my estimates versus how long something actually takes me.

Anyway, thanks for the engaging read.
Jared