I worked for a social media management app company (Sprinklr/Hootsuite/Sprout-esque) and we had the exact same problem. My ballpark estimate is that at least 90% of the messages that went through our system were extremely obvious spam or scams. However, there was a huge focus on MAU growth, so there was no way in hell we were going to attempt to combat this.

For legacy implementation reasons, the backend service that published messages to Twitter would lose 1-2 queued messages every time the service was redeployed. Bleakly, the engineering team decided to not consider this a major issue, despite it being a core competency of the platform. It was simply extremely unlikely anything of value was being lost.

Did you try to introduce a new measure "MAU excluding bots"? Then over time the company could refocus around that and better policies be made

That seems like a good idea, but I suspect introducing a "MAU excluding bots" metric would look to investors like a sudden drop in MAU by 90%.

It would seem reasonable for large investors, -those who have influence on a company, to require MAU excluding bots figures as part of their due diligence. On the other hand: FOMO, guess.

The only economic actor directly hurt by bots is the advertiser, right? To some degree, it's against investors' incentives to attempt to strictly exclude bots.

Obviously in the longterm it hurts value, but "doing things that don't scale" creates a lot of degrees of freedom (user growth, cashflow, funding) that you can use to pay off your "tech debt".

Thanks to this thread I now see why the bot problem is sad difficult to solve. It sticks deep. Damn.

I wasn't nearly senior enough to have any influence in something like that, but in the internal culture, MAU was seen as such a strong proxy for success that suggesting the metric was flawed would have been sacrilege. The company had big parties and put out press releases every time a multiple of 1 million MAU was hit. To point out that most of those users were bots abusing our service's free tier would be noticing the emperor had no clothes on.

Then more bots come in undetected and you have to rebase again when you find them. It's a continuous process. You need very understanding investors to pull this off. At the same time there's only so long you can sweep inauthentic engagement under the rug until someone calls you out for it.

I don't understand why investors don't ask for numbers that exclude bots. Or maybe they do, and we just don't hear about it? This isn't some unknown problem, right? Bots have been around forever and they are a problem in many platforms, not just Twitter-like platforms.

The core problem them was a lack of integrity on the part of the executives. Which I am not saying they were especially unethical. Lack of integrity is very common in executives. But still, that is the root cause there.

Is it possible that Facebook’s apps could also be using bots to boost their DAUs etc?

This nailed it. Public companies have to "grow" and if growth is bot-driven then combating it is not goin to happen. By taking the company private Elon could (in theory) throw out MAU metrics, engagement metrics, etc.

He could "Make Twitter Great Again" by killing the bots, and rage-engagement, and making Twitter interesting and fun again. If that raises Twitters value by 10% he gets a $4.5B return.

Everything I've read says he's taking it private. Yes.

Doesn't advertising revenue require the same metric to be high? Assuming he wants to make money or just be self-sufficient, not operate at too much of a loss.

If you can improve the general signal/noise ratio I think that'd be big for advertisers. I think it's one of snap's strengths actually.

If Elon makes it subscription based in addition to being private, I’m confident that the bot problem will be solved.

Publicly traded companies are too often affected by short sighted investors.