Not only enormous global businesses, but a completely different type of platform.
https://meta.discourse.org/t/give-discourse-social-media-look/23711/29?u=purldator
https://meta.discourse.org/t/give-discourse-social-media-look/23711/32?u=purldator
https://meta.discourse.org/t/give-discourse-social-media-look/23711/62?u=purldator
Jeff - I never said it was easy, it just seems that if Facebook can invest .16% of its time (20 engineers and data scientists out of 12,000 employees) working on their feed algorithm - then perhaps Discourse could invest something like that to create at least a rudimentary algorithm to do some sort of prioritization.
And, Iâm sure youâre an Eric Reis fan - it seems you could start with some sort of minimum viable algorithm. I suspect (correct me if Iâm wrong) that any simple algorithm that even just takes into consideration 1 factor - the titles of the topics that the user has âlikedâ - would be better than just a chronological firehose feed of all the data / messages / topics.
What Iâm saying is the following:
Investment could start low with a minimum viable algorithm using a contract data scientist (Stanford and Berkelely have lots of these people, and if you combine the data from all your hosted forums, you have some good data to evaluate). Perhaps get some PHD students for cheap at UC Berkeley to do this for free or near free as a PHD thesis.
Constrain the number of variables initially - just try to get a prioritization of the Topics (not the individual messages - since really each topic is one conversation).
I may be wrong on this - but it seems that virtually any effort at prioritization is better than no prioritization (other than time).
I mean - you donât have all your engineers just randomly work on possible projects for Discourse right? Why do we expect our forum users - who are likely just as pressed for time as you guys - to engage in randomly prioritized topics?
Seems it would be easy enough to create a custom badge that has a nag type of descriiption. eg.
Cat Liker - Youâve liked n number of posts in the Cat category. Have you considered Watching it?
They both provide a flow of social news and information to a set of interested users, in a centralized Feed.
And a key problem in both of these platforms is that you have to provide a high signal to noise ratio of information or you start to lose userâs interest and visits - and that is key to the survival of your forum / or platform.
There is enough of a similarity that it seems people might want to try to learn how one has overcome the issues using algorithm-driven customization (over time - this isnât a binary issue).
Part of the issue in discussing this here is that I suspect that perhaps 95% of Discourse forums get less than 300 user messages a day - the point at which perhaps this has started to become an issue in my forums.
Reminder that these 20 engineers and data scientists are also supported by the rest of the Facebook organization. They do not work alone, in isolation, on systems they provided for themselves with data that they gathered on their own.
Since our entire company is 9 peopleâŚ
That does not support your argument stating Facebook data is relevant and can be used to gauge Discourseâs performance regarding user retention.
Users log into your community to engage a specific topic.
Users log into Facebook to post a status message, post photos, chat about anything, play games, and also for SSO purposes from other websites asking for authentication.
If facebook and many other social / information oriented platforms can crack the retention puzzle - I have full confidence that the Discourse team can do it too; theyâve built a great platform and this would take them to the next level. If they can - there is no question in my mind that they could completely dominate the discussion market (and not just do reasonably well, as they currently are)
as they say:
âNews Feed is at the epicenter of Facebookâs success. Over the past nine years, the product, which was initially controversial, has evolved into the most valuable billboard on Earthâfor brands, for publishers, for celebrities and for the rest of us. For years, the News Feed has been fueled by automated software that tracks each userâs actions to serve them the posts theyâre most likely to engage with. That proved successful in helping News Feed generate more revenue for Facebook than any other part of the site.â
I agree.
But you cannot use Facebook data as evidence to support your stance. This is comparing oranges to hammers. I understand you really believe this feature will help Discourse. This is me wanting to help you make a better presentation to that belief, for your benefit.
I suggest employing a different set of data not related to large corporate businesses and/or social media sites.
The issue is that this type of user-driven customization only happens at the social platform level - individual forums / small medical markets are never going to justify this type of unique investment - the numbers are way too low.
And ultimately - I donât believe the behavioral psychology fundamentals that are driving this for the broader population are any different for my subgroup. This is a human issue - people donât like ânoiseâ - and at some point when the noise level becomes to large they tune out.
Thatâs fine, then work on reducing the noise on your site. You can do this by posting a meta topic about what ânoiseâ is and means for your community, PMâing users who generate noise, or ultimately moving them away from your platform.
Wow, sorry, I thought you were asking for additional ideas on âengaging user participationâ. For what itâs worth, those who are offering additional ideas in this topic that arenât algorithm-based are, as far as I can see, successful & award-winning community professionals with track records of building and sustaining huge online communities. 
For the others participating in this topic who are interested in discussing such ideas, remember that there is actually a science to community management, and (IMHO) it begins (before building or changing any software) and ends with human engagement, which can be measured, quantified, and tracked over time. (The most successful community managers do this.) If there isnât compelling, trustworthy, interesting content for a majority of people to keep coming back, they wonât. 
Social networks are popular because those compelling reasons (friends, colleagues, favorite brands) are there. The expensive algorithms mentioned here only amplify (hopefully, maybe) the right stuff. If there was no compelling content to be amplified, Facebook would become a ghost town just as MySpace and others before them.
An idea:
Moved to new topic because this thread is hard to follow
Played around a little with smoothing today, in case anyone wants to explore further:
var smooth = function(data, windowSize) {
var smoothed = [];
var movingWindow = [];
for(var i = 0; i < windowSize; i++) {
movingWindow.push(data[i]);
}
for(var i = 0; i < data.length; i++) {
movingWindow.push(data[i]);
movingWindow.shift();
smoothed.push(movingWindow.reduce((a, v) => a + v, 0) / windowSize);
}
return smoothed;
}
var samples = rawData.map(r => r.y);
var data = {
labels: rawData.map(r => r.x),
datasets: [{
data: samples,
label: model.get('title'),
backgroundColor: "rgba(200,220,240,0.3)",
borderColor: "#08C"
},{
data: smooth(samples, 7),
label: '7 Day Moving Average',
backgroundColor: "rgba(0,0,0,0)",
borderColor: "#C80"
}]
};
This isnât always bad, so long as it doesnât just become nothing but spam. If itâs good discussions then it can bring members closer and make them more loyal to the site. Iâve seen this time and time again on larger forums, they have places to unwind that are off-topic for the siteâs niche, the relationships made there greatly increase engagement.
When users make friends on the site they tend to spike up to the top 1% of active community members, coming on often and becoming better contributors overall.
That still pushes them outside of the site, back to their email app, then back to the site. maybe just some sort of mobile banner that has the login already pulled up so they can enter their information, maybe allow users to have a PIN for when on mobile that they can set themselves that would speed up login.
This. Unless the platform can read and manipulate your mind as you visit the site, itâll never be perfect. You could love certain topics and not be in the mood for them when youâre on the site that day. And if anyone ever found the solution to that problem theyâd take over Facebook and all other social media sites or get paid a lot of money by one of them. Itâs a HUGE ask of Discourse to be able to do this when the king of social media (Facebook) still canât.
Emails donât have much of a high tolerance rate and itâs far more common for people to ignore it than pay attention to it.
âAccording to ZipStripe research, it takes a recipient an average 6.5 hours to view an e-mail, but only 15 minutes to view SMSs and push notifications.â
By the time people open the email, if they do at all, the conversation will likely have died down. Getting users to engage quickly with content will keep the conversation going, keep the other parties involved as well, thus furthering the discussion since the people that are part of it will all be notified at the same time via a method theyâll actually notice quickly.
A relevant tangent was discussed a bit ago, with using trust levels to ârewardâ users who post to particular categories, or not reward at all when posting to other categories.
https://meta.discourse.org/t/disable-categories-topic-from-the-user-statistics/34717/6?u=purldator
General discussion boards and participating on them would offer no in-system incentive with this idea of mine. But these off-topic boards do offer a needed psychological-related outlet; one I feel every niche discussion community should have if it wants to survive for very long.
This may be part of a sound remuneration strategy: âWhy should I post on-topic and offer free original content for google to attribute to your website. Give me incentive to stay. Oh? I can make friends who be as passionate as I on this same niche topic, and talk off-topic on other subjects to mix things up a bit? Iâm game.â
One problem I notice, is that users arenât reading their messages. Maybe make the notification bubble for unread messages to remain up there until the messages are read (not just the notifications)? I send a lot of personal welcome messages and other things that would drive engagement up and satisfaction with discourse, but if these messages arenât being read, then thatâs futile.
Another thing I notice is that some users try to reply to digest emails⌠anyone else notice this?
Just a few things on top of my head. I am pondering whether at some point I want to switch back to the category as main entrypoint of my community because that would force folks to make a few clicks. The situation I am observing right now looks a bit like people lose interest in the forums to some extent because with the default few they see very plainly how little is actually going on sometimes, whereas before, the inactivity wasnât so much in plain sight, and users just kept clicking through to what they were interested in and then started posted nonethelessâŚ
I believe if itâs a private message then it stays until itâs been read. At least thatâs my experience over the last week.
You are right, the bubble remains, but what if there are other notifications, then maybe a not well-informed user will see the bubble, click on his avatar (if he ever set one, another problem), but then not see a notification in the list that pertains to the bubble (from his/her perspective)
In this example (my own), I see a message at the top of the list. What would even be better, is if the envelope symbol (slighly yellow because I moused over it) would somehow change its look as long as there are unread messages, I wonder what @eviltrout thinks about that⌠might be helpful?
Totally agree on this! Bring members closer and help them becoming friends is always a good thing 
