Sign Off - Don't get me started

JM&J but I hate the term “Sign Off”.

If you want me to sign off on the software, give me complete control and then I’ll sign off…

I’ll provide data to help make a decision, but I won’t sign off a release/application/system etc.

Its not that I am avoiding any responsibility or culpability for my actions, but I won’t and cannot assume responsibility for all that has come before me.

I apply my skills/talent/deviousness to figuring out if we’ve done what we set out to do and if we’ve managed to stop people from breaking it afterwards. I want to ship the release, but I also want to make sure I know as much as I can about it before it goes out.

I’ll share with you all of that information/knowledge/insight so that I can to help us make an informed decision as to whether or not the release should go ahead.

But. I. Will. NOT! Sign. Off. The. Bloody. Release.


My last company had this ritual where a Senior QA had to sign off on the work, there were two of us but after a bad miss they insisted that I had to approve all releases, it was not a fun time and in the end refused to sign off on another’s work unless I got to test it myself. They didn’t seem to care what happened as long as the ritual was observed.

At the CCB meeting I would get grilled on inconsequential items, while they would pass items with identified risks in minutes. E.G., the bits they could understand versus the bits they couldn’t. Still they had to have a signature. The good news in this is of the many I told you so’s were all approved risks which is the best I could wring out of that project

1 Like

You’re right- “signing off” a release has become a distinct tick-box exercise in many organisations.
Surely, if using modern tracking tools, you can see that all steps required in the team’s Definition of Done are completed, everyone has approved it as per the team’s requirements for release, so the release is good to go.
A separate sign-off is just adding overheads to the work required.

As devil’s advocate, I can see where monolithic organisations require a sign-off, but I do think that changing outlooks on approvals should be encouraged.

1 Like

There are so many reasons to hate sign-offs. It’s definitely not agile - violating Customer collaboration over contract negotiation. It shows a sign of distrust. Prove to us that you really tested it - and it would have to be all tested in the final build if you really wanted to be sure.

Having QA sign off, also shows that they believe quality isn’t everybody’s responsibility, just QA’s responsibility. Does QA have the power to change the development process? No. Does QA have the authority to demand that Devs write, run and share the results of their Unit Tests? No.

It’s also a CYA mechanism. If something fails, we will blame QA. You will also most likely find that even though you must sign off, you do not have the power to block the release. I remember reading a blog of a QA Lead that tried blocking the release and he was just overridden.

Ah, the checklist (tick marks). A key feature of many large organizations. Checklists are good to make sure you didn’t miss anything, but I worked for an international company and the Test Plan document was this first item on the checklist. Until it was completed, everyone on the project acted like it was the end of the world. The reality is that the test plan wasn’t followed if the project fell behind. The QA Lead (me) had no authority to enforce it. The PM ran the show. To recap: creating and getting sign off on the Test Plan was critical, but it didn’t have to be followed and no one was very interested in the actual quality of the software.

If there is to be a sign off, why not have the Business (end users) sign off on it?


Is this coincidence? or fate? I knew I’m not alone on this :wink:. A few days ago, I tweeted about What will be the three things in testing practices we can change to bring a brighter future in testing? By reaching out to leaders and awesome testers. []

The response was amazing and prompt me the idea about I should write a blog post to give some context.

Today I was trying to decide where I am going to write it - MoT Club came up straight away, as I’m too lazy to maintain my own :laughing:. Nah, my WordPress probably smells since I haven’t looked after it since I’ve created it :joy:.

Then, I found this blog post. I think I should reenforce it by another blog post :thinking::grin:

Testing can only show how a product behaves under certain circumstances.
It cannot show a products works per se because:

1 - it works is subjective;
2 - Due (1), it can evaluated through an infinitude of perspectives.