Wednesday, August 1, 2018

Seeing what is Right when Everyone Seems Happy with Wrong

I'm uncomfortable. I've been postponing taking this topic up in a blog post, convinced that there must be something I'm missing. But it may be that there is something I'm seeing that many others are not seeing.

A few months ago, I volunteered at work to try out crowdtesting on my product. It wasn't the first try of crowdsourcing, quite the contrary as I learned we were a happy user of one of those services for some other products.

The concern I had surprised everyone, providing a perspective no one else seemed to think about. I was concerned if using a service like this would fit my ideas of what is ethical. I was not focused on the technicalities and value and usefulness of the service but whether using the service was right. It isn't obviously wrong because it is legal. Let's look a little more where I come from.

I live in Finland, where we have this very strongly held idea that work needs to be paid for. Simultaneously, we are a promised land of non-profit organizations that run on volunteer work having  huge numbers of organizations like that in relation to number of people. But the overall structure of how things are set up is that you don't have free labor in companies.

The unemployed and the trainees get paid for the work they are doing. And this forces companies to be on good behavior in the social structures and not leech of the less fortunate.

So if my company in general wants someone to do testing, they have three options.

1) Customers Test
They mask it as "no need to test" and force their end users to test, and pay for structures that enable them to figure out what a normal user is actually complaining on, and structures for fixing things fast when the problems hit someone who not only raises their voice but walks out with a relevant amount of money.

2) Someone in their employee list tests
They pay someone for some of the testing, and some companies pay testers to get more of the whole of testing done. If you tested half of what you needed, you still tested. And this is the narrative that makes the current programmers test it all trend discussions so difficult in practice. There's testing and there's testing. It takes a bit of understanding to tell those two apart.

3) Someone in a service provider organization tests
They pay for a vendor for doing some testing. The vendor is an abstraction, a bubble where you conveniently allocate some of your responsibilities, and in many cases you can choose to close your eyes at the interface level.

Crowdtesting belongs to the third option, and relies in my view on closing your eyes at the interface level.  Crowdtesting is the idea of paying a company for a service of them finding you a crowd. They find the crowd by paying those testers, but I'm not convinced that the models of paying represent what I would consider fair, right and ethical.

So, I pay a legitimate company 5000 euros for them to do some testing we agree on under the label "crowdtesting". Yay! Is that really enough of thinking from my part? You get 30 testers with that, so cheap! (numbers are in scale, but not actuals). They even promise most of them will be in Finland and other Nordic countries. If your alarm bells aren't ringing, you are not thinking.

The more traditional companies producing things like coffee or clothes know painfully well it isn't. You really don't want your fashion brand be associated with child labor, inhumane working conditions or anything other nasty like that. Surely you can save loads of money using companies where their delivery chain hides the unethical nature of how the service is created, but you risk a reputation hits. Because someone hopefully cares.

Crowdtesting companies are not in the business of child labor, at least to my knowledge. But they are in the business of forcing my people (testers) away from being paid for their work, to be paid for their results. And with results being bugs and checkmarks in test cases, it's actively generating a place where folks in these programs are not necessarily paid for the work they do.

The way the monthly fee is set up for the customer makes this worse. There's a limit on how much of these paid checkmarks you can allocate a month, but you're told you're allowed to ask for unlimited exploratory testing  on top of that. The financial downside for the people doing testing in the low end of this power dynamic is that they are only paid for bugs customer accepts.

Many people seem really happy with doing testing in these schemes.

The ones promoted to the middle management layer get paid for their hours, leaving even less of the financial cake for the low end of the power dynamic. But why would they care, they get paid.

The ones privileged enough to not really need a job that pays you a salary get paid whatever, but since the money never mattered, why should they care? This is a hobby on the side anyway.

The ones living in cheaper countries getting paid the same amount as the testers from the Nordics may actually make a decent amount of money in finding problems, and might not even have a choice.

The really good ones who always find problems can also get paid, as long as the effort of finding something relevant isn't getting too high.

But the ethical aspect that I look at is local. The low end of the power dynamic are the testers kicked out of their projects who no longer need them, but actually do. They just no longer want to pay for all of their work and hide this fact in the supply chain.

Maybe I should just think more global. I don't know. But I am uncomfortable enough to step away. I don't want to do this.

With the same investment, I can get someone who works close to the team and provides more value. Without being exploited. With possibilities of learning and growing. Being appreciated.

And I say this looking from a higher end of that dynamic. I'm not losing my job in this shift.

I don't like crowdsourcing. Do you?


1 comment:

  1. I don't like crowdsourcing either, but for a different reason.

    With any model, you don't have control of the expertise of the people doing the testing. Plus there's next to zero domain knowledge about the product and company, and who knows how much they know about the sector the software is going to operate in.

    It seems like more of a "warm fuzzy" idea that's meant to check off the "yep it's tested" box, and keep costs low.

    ReplyDelete