Exactly. It took me a while to understand that this article was written for community managers of brands, not people developing social networks. They keep saying engagement has no value, but to social networks and to "influencers" who both make their money off of ads, that is exactly what has value.

> the reason "engagement" is the metric is because that's what platforms monetize

...and the implication of that is that if we change the monetization model, then the metrics also change.

For a company selling a paid product, metrics become things like the effectiveness of the sales funnel, how much support the average user requires, the rating of the product on app stores, etc - which are not ideal, but still far better than "engagement".

You might be surprised to discover that even for paid products, engagement is often still the #1 metric.

Simply because people who are engaged with a product/app/whatever are by far the most likely to be repeat customers -- to renew their subscription, to buy the new version, to adopt the next product.

Once engagement drops, future revenue dries up. A drop in engagement is the canary in the coal mine -- it shows where your business is going wrong months/years before it shows up in sales stats or renewal rates.

So actually, the metric doesn't change at all. E.g. Netflix is obsessed with engagement even though it's subscription-based rather than ad-based. Because guess who winds up cancelling their subscriptions?

s/surprised/disappointed/

A user's engagement with your product is a great metric, because what you value is users using your product.

But for a COMMUNITY it's the wrong metric, because you don't value negative comments in your Slack no matter how many there are (in fact, the more there are the worse it is).

>> Good article, though I think it missed explicitly making the point that the reason "engagement" is the metric is because that's what platforms monetize.

In the case of "community" around a product, the purpose of the community is usually to provide free support. That's not so much monetization as it is cost savings. For that to work, people need to be "engaged" or kept around so they are present to field those support issues.

Engagement is still the wrong metric for this purpose, because flamewars rate much higher than single comprehensive answers. In fact, partial or incorrect answer rate better, because the questioner will have to come back or someone will need to post corrections or clarifications.

Engagement is a better proxy for value destruction, if you define people's time and sanity as valuable. (It's not perfect; a thriving community will rate highly.)

Engagement is just a more sophisticated version of 'eyeballs'.

For community building you want successful events, you want interesting questions to be asked, and interesting answers to be offered. Someone could write a book on how to make communities work. Probably several books.

There are several books on precisely this.

You are completely right on Goodhart's Law. I know this from experience.

I put the customer first in the belief that 'the rain follows the plough' and, if you put the customer first, then all else follows.

I see metrics as there to be cracked. If you are measuring the right things then you can see where the problems are, fix the problems and move on to measuring different metrics that are more specific and focused on the remaining problem areas. For me it is a campaign where initial reactive firefighting gives way to calm, proactive problem solving, fixing problems before they are of consequence. Along the way many unexpected things are learned about the customer.

I do all this with a focus on the customer, meanwhile my colleagues in marketing are out for themselves. They are the dead weight keeping progress back as they have to do things like buy traffic and hold on to keyword optimised URLs that are cargo cult SEO. Their reports for their micromanagers are always measuring the same stuff, e.g. 'engagement', yet they could be engaging customers solely for the purpose of pissing them off for them to never come back.

My favourite is the pop up shown as someone leaves a website begging them to stay, or the sign up to the newsletter popup. These could get one extra sale at the cost of pissing off a million people but if you only measure the former then it all seems good.

In the capitalist world only one metric matters to the shareholders. Once I worked for a highly successful company that sent out sales newsletters with no measurement whatsoever of open rates or click throughs. None of that was needed. The problem was emptying the warehouse too quickly to have no stock left. Customers would be waiting for our email, not ignoring them. We had plenty of engagement with the customers as they actually bought stuff.

Subsequently I have worked for agencies where they spend a fortune on some email service and they get these fancy graphs to show the clients of these open rates the world over. Which had wow value at the time and 'engagement' stats. But I had come from somewhere where we had no need to measure that stuff, we weren't looking at these charts, we were coping with a deluge of orders and had no time for that.