What is the most prevalent technique for estimating test efforts?

Replies to This Discussion

Permalink Reply by Michael Bolton on January 30, 2010 at 9:40am

I wrote a tool that does just that and have been using it for years now and we come within a day of our estimates.

Suppose I have an estimate such that I'll stop testing on Wednesday at 4:00pm. On Wednesday at 4:00pm, I decide to stop testing. Mind you, I could decide to keep testing until the end of tomorrow, or I could have decided yesterday to stop testing. The point is, if I'm involved in the decision, was it an accurate prediction or a self-fulfilling prophecy?

---Michael B.

▶ Reply to This

Permalink Reply by Paul Victor on February 1, 2010 at 9:55pm

I like the way you think... however based on the test plan and milestones set within that test plan the tool has been very usefull in creating the guideline for being test complete. There are always other variables that could determine ending a testing cycle, that said this tool has been used project after project and multiple companies and have proven to work well in forcasting test hours.

-- Paul

▶ Reply to This

Permalink Reply by Michael Bolton on February 13, 2010 at 2:54am

I've noticed that whenever I've asked someone the question that I asked above, I've always g0t an assertion that the tool was useful, but I've never, ever got an answer to the question. I should start forecasting that.

---Michael B.

▶ Reply to This

Permalink Reply by Srinath on February 10, 2010 at 10:58pm

I have used PERT-DELPHI & UseCasePoints for estimating some of my test projects.

But we really need to look into how many projects, we do get to Estimate Independently for Test Activity. Frankly speaking not more than 3 out of 10 test assignments, would a Test Manager have the opportunity for Estimation. Most others would be based on the Schedule & Deadlines of the Project (Analysis, Architecture, Development, Test & Deliver).

Also another Important thing which would be a key driver for an Opportunity, would be the type of Testing. Usually a UAT (done by third party / Independent Testing Team) would score over IT & ST (Integration & System).

So if we need to grab those Opportunities, we also need to emphasize on getting the right kind of Project.

Test Projects, which would not provide an opportunity with "Test Team provided & Accepted Estimation" are usually STRESSFUL.

As far as WBS method of Estimation is concerned, I am not sure if it is substantiated with appropriate Metrics (for decision making). The WBS would again be derived from Development activities, and may not be the right Approach for Testing.

▶ Reply to This

• ‹ Previous
• 1
• 2
• Next ›
• Page