Justin Isaf AMA

Hi Neeshu, Joe is right, and Richard (the guy behind Feverbee) has a great book called "buzzing communities" that really lays out a great path to starting successfully. It's a lot of hard work, generally takes about 6-12 months and doesn't look like much is moving.

It is definitely a case of "if you build it, they probably wont come" so start with the platform last. Find great people, get to know them in your network, then bring them together all at once to get critical mass - this might be on a blog, in an email chain, maybe a twitter chat, or something simple where you're the primary content creator and driver. Then consistency is key, make sure you're consistent about starting conversations. Even if they are a month apart, people should know that the 1st of the month is when they all gather to chat.

And then always make sure that you're making them more awesome (again, look up the "minimum badass user")

Good luck!

It's been fun! Any ping me on Twitter any time you have questions that don't get answered today! @justinisaf

Great answer!

Great questions.

1) Most definitely the hardest thing a community manager will ever have to do is to build critical mass - until you have users, nobody sticks around, so you don't have users, so nobody sticks around. The Dribbble/Medium model isn't actually about exclusivity, although it does accomplish that. It's about creating a sense of scarcity in a world of infinite availability and leveraging networks in a really smart way. Without going too deeply into game dynamics (which is a MUCH longer post), by creating a sense of exclusivity, you make everyone "aspire" to be a member of your community, thus more people try to join. But you're also creating recognition for your core users by giving them control over a valuable resource. If you give me an early invite to something it says 4 things: 1) you're cooler than me because you had it first, 2) you think I'm worthy enough to be invited which sociology has proven makes me think you are worthy enough for me to look up to (making both of us more awesome in our own eyes), 3) makes me owe you for giving me this scarce resource so I am less likely to "bounce" out of the community once I sign up and try things and 4) it says that you're a member of my network and I'm a member of yours (both to each other and to the system) which helps the community manager to more quickly and easily build a more dense network.

2) This is entirely a function of your community. Different communities react to different things in different ways. At the end of it though, you want to make your users more awesome. Everything is just a gimmick if it doesn't start with what makes your users more awesome. Look up Kathy Sierra and her "Minimum Badass User" talks.

Patience and that it's just the internet. No matter how bad it seems, it's not as bad as it seems :)

#1 I'll second Tim's response below for #1. Empower your community members to be awesome, and they will run your community for you. It's the "first followers" who define your community anyway, so let them grow and run it. If you haven't seen it yet, watch Derek Sivers' TED talk (http://www.ted.com/talks/derek_sivers_how_to_start_a_movement.html), and focus on the few first followers and primary advocates. Let them worry about everyone else. If you're trying to talk to everyone, you're a network manager. And that's just silly. The job of a community manager is to put themselves out of a job because their communities run themselves.

#2 Sad truth, I don't have any. There's a saying that "those who work in the community industry, don't commune in the industry in which they work". I spend all day dealing with people shouting at each other on the internet. When I'm not working on my community, I go outside :) BUT, check out http://thecommunitymanager.com/, http://mycmgr.com,http://www.communityroundtable.com an go from there :)

#3 Bad-ass community managers generally aren't trained, they just are...the training is in how to apply their skills to an online space. If I wanted to learn from scratch all over again I'd go to work at a summer camp. I've said it before and I'll say it again, if you can convince angsty teenagers to become a cult in a couple of months or less, you can run a community.

#4 As a true community manager, I'd have to say I'd try to convince the 100 duck sized horses to fight the one horse sized duck for me. I don't do anything myself. Puh-leeze.

I'd say go to communl.com and shoot me an email ;)

This is THE question. If anyone figures out an answer that consistently works please let me know :)

There are a few ways to approach this. In an "owned community" (i.e. one on your site, not on facebook) the easiest way is to work in some form of "cost deflection". Support communities create content that deflects call center or support desk inquiries and thus reduces staffing needs. I once ran a community that did work that would cost us $2-$4MM a year to do, and I was the only paid staff working on it, and I only worked on it 10-15 hours a week...we calculated that my time there had an 8000% ROI. This way you don't need to justify value added, just pure cost deflection. But, that often isn't the most convincing strategy long term, because it's a "negative" justification not an "additive" justification.

The philosophical way of approaching it is that "everything is a community" - your familiy, your co-workers, this AMA, everything is just a system for people to interact with other people and therefore you should be the center of that interaction in your interest space. Even something as mundane and "secondary identity" as invoicing has had a community built around it (see Freshbooks.com, amazing stuff over there).

The social media way to answer this was "fuck communities, let's just get links. Easier, cheaper, more measurable". And the sad thing is, it's true, which is why we've seen a lot of moves away from communities and into networks (see my answer to Sarah K above for the difference). I can measure likes, followers and clickthroughs which make that much more justifiable. But the irony to me is that every one of those clickthroughs and followers requires work on your part. Everything is worked for. With communities, they do the work for you. It's free and infinitely scalable once you set it up because every node enables exponentially more nodes to fit in, vs a network which only has one node and that node is you. Take you out of your twitter stream and the twitter stream dies. Take you out of your community and it will live on.

The best way to do this in every case is to make up a number. Any number. Measure anything. Signups, comments, visitors per day, "the ballyhoo co-efficient" (whatever you define that to be) and report it. Community tends to be this thing that lives off in the corner somewhere and everyone can ignore it. By giving it a reported number, people psychologically assume that it's something that they are being measured against and everyone is more likely to jump in and make it go up and to the right, including management.

The platform is the wrong thing to be focusing on first. You can build a highly engaged community with email.

Ask yourself what makes your community more awesome, define the features you NEED to do that, then find the platform that has those.

Bells and whisles are great, real time is fine, but at the end of the day you need comments, notifications and some form of friending/following and that's really it. Everything else will make life easier (or more complicated) but it's just extra stuff.

Focus on the people first and the platform wont really matter.

And because videos on YouTube tend to be so personal, the attacks are personal as well. They aren't targeted at a person (Bob27) but at me, Justin, and my poor dancing skills...and that hurts.

The number of times that bullying rears its ugly head...it's scary that there isn't more done.

Yeah, especially as communities grow (and network density tends to fall), groups become vital not only for growing engagement but for sheer survival.

By allowing sub-groups of a main community, you not only give people the chance to create denser networks within the larger community and give them a specific identity, but if you also give them some degree of control over that community, the recognition you're providing can tie the leaders of that group to your community for a long long time.

It's a great way to go at scale. But be careful of fragmentation when you're small because more spaces makes it harder to create critical mass in any one of them (again see the empty restaurant syndrome).

ooh, interesting question :)

First, and always, try to get them out of public sight as quickly as possible - that guy tends to be self righteous and putting them on a public defensive will only make things impossible to deal with. It also gives them less chance to amass a following. A small, vocal, group of people will tend rally around that guy for the lulz of it, or because they have some legitimate complaint that you haven't addressed, but when fueled by that guy, they will also become impossible to deal with.

Avoid accusing them of anything off the bat. It just puts people on the defensive. Try wording things as misunderstanding and talk it out.

Then everything depends on who it is (see http://www.flamewarriorsguide.com/). Sometimes you can reason with them and it was a genuine misunderstanding so you can engage them and turn them to great posters. Other times you need to cut ties quickly and effectively, and hope they don't become whac-a-moles that create 100 new accounts to troll you.

Generally just remember that there is a real person on their side of the screen. Treat them as humans, and they will hopefully realize that you're just a person too and interact with you on that level.

If you have an example that's more specific, we can go deeper into specific tactics.

Haha, I'd cheer, then I'd cry, and then I'd moderate the shit out of that thing - it would only cost $5-$10 million a year :)

Seriously though, I understand their reluctance to moderate. There's a lot of liability, a TON of content and not a lot of incentive to do it.

BUT, with their attempts to get content creators on there, and to create followings for them, they do a genuinely horrible job of incentivizing the creators to engage with their community. The comment sections are the pits of hell, and the tools aren't great for the type of conversations that might really lead to something special.

But then, I don't use YouTube a lot, so I might be missing something...

Any additions you'd make to it Robert? How would you measure your success in a community?

Hey Jenn!

My typical day at HuffPost is not a typical day for a CM, to be sure. I was solidly in middle management, overseeing roughly 90 people spread across 9 or 10 different teams. My day was frantically answering emails in between back to back product meetings and fire drills in the community.

I loved it :)

Generally it would start around 8 or 8:30 with checking the morning emails, the daily stats from the day before, and dealing with anything urgent. I'd work from home until around 11 if I didn't have any meetings (ha!) and then would be at the office until anywhere from 6 until whenever (I spent more than one night at the office when things were afoot - Osama Bin Laden, Elections, days when our machines decided to take a vacation, things like that).

I had a desk, I don't think many people knew that because I was never there because of meetings and whiteboarding sessions. Lunch was always a working lunch. Dinner was often eaten out or taken in because I was too tired to cook (thank goodness for Teng Dragon...open until midnight on weekdays, 2am on Friday and Saturday).

I was on-call 24/7 so I got lots of calls at ungodly hours of the day...funny thing about internet commenters...they never stop commenting.

I worked with amazing teams and we did amazing things and it was amazing...and lots and lots of work.

Definitely try to track down Richard then, he's one of the leading thinkers in the world on community management and is based in London.

What my superiors measured us on and what I measured us on were different.

The bosses wanted to know comments per day, signups per day and cost per comment moderated. The top line stuff that is easily reported and easily understood.

I measured us on other things, not all of which I can go into too much depth on sadly, but in general, thread depth, the ratio of replies to original comments, comments per hour processed and accuracy of that processing, flagging rates per vertical, distribution of comments both by creator and by vertical, comments per article, the rate of growth of commenting vs page views, and the average shortest path between two users. All of these were measured over time. There are no absolutes for what is "good" in communities because they are so diverse. I didn't measure anything "Social Media" because that was a completely separate team.

There's a lot I'd like to have measured, but didn't have the resources to implement, and that's always the story of it :) If I had my ideal, this would be my metric: http://thecommunitymanager.com/2012/10/25/how-to-use-the-network-density-formula-to-measure-the-health-of-a-community/

Oh but if you want to look at the leaders - for machine moderation: Crisp Thinking from the UK, KeepCon from Argentina. For human, eModeration from the UK and ICUC from Canada.

I'm so glad you asked about machine moderation, one of my favorite subjects to nerd out on :)

Machines (alone) suck at moderating for exactly the reason you outlined. They tend to be about 70% (+/-10%) accurate, but sometime run into the False Positive Paradox (http://en.wikipedia.org/wiki/False_positive_paradox) which makes accuracy significantly worse.

On the other hand, humans (alone) also suck - they are slower, more expensive, have non-systematic biases that are hard to test for, and come with baggage in the eyes of the community - they tend to be about 90% (+/-5%) accurate.

There are no systems that I am aware of that solve this perfectly. I would argue that what we did at HuffPost was pretty darn close solution - by combining the best of both and having both admit their weaknesses, we created a system where the machines acted on everything that they were sure of, but if they were unsure, they would pass the comment on to a human team to check. By combining the best of both worlds, we could get closer to "right" at significantly lower costs. 

It may not have been the best community on the internet, but it allowed us to be the least bad at scale...and at 400,000 pre-moderated comments per day, that's not too shabby.

Yeah, I tend to delete discussions about moderation pretty quickly (they are always off topic, and see the Broken Windows Theory) and try to take the conversation into private where it's easier to have a reasonable discussion without anyone feeling like they have to be on the defensive and lash out.