Most social groups have a big retention problem. People join, participate, and leave. They use the group solely as a place to get free advice. If they don’t need free advice, they don’t visit.
Some retention rates are shockingly bad. We’ve seen rates as low as 1 in 1500 for newcomers*.
The average for online groups hovers around 1 in 50.
That’s just 1 member active after six months for every 50 that join. To reach 1000 active members, you would need 50,000 registered members. Let’s assume your newcomer to registration ratio is 1 in 5 (high, but possible). You would need to attract 250,000 unique newcomers to get 50k registered members.
Getting 250,000 newcomers isn’t easy for a company without a huge existing audience. For comparison, a promoted tweet costs $0.2 – $4 per engagement. That’s the equivalent of $50k – $1m of work**. And this is just to reach 1000 active members.
If you’re selling low-margin products to the masses, 1k active members won’t cover the costs of creating an active member. For most groups, 1k members alone isn’t going to have a big impact.
If you can’t improve your retention you’re not going to survive over the long-term.
The Biggest and Best Investment of Your Time
[tweet_dis]The biggest and best investment you can make in your group today is on retention.[/tweet_dis] That means changing how members see the community from a free resource where they can get advice to the place where people like them hang out.
You can’t do this by upping the informational value. Giving away more free advice to people who only visit when they want free advice isn’t the solution. It doesn’t get members to visit more frequently in the first place.
The best (and possibly only) way to increase retention is through cultural assimilation of newcomers.
If someone isn’t assimilated within a group’s culture, they will never see the community as a place where they can satisfy their social needs. They will never feel it’s part of their identity and their social group.
4 Metrics To Measure Cultural Assimilation
Sociologists measure cultural assimilation along 4 lines; socioeconomic status, geographic distribution, language attainment, and intermarriage.
We can adapt (and re-order) this slightly for our purposes.
We track intermixing, skill/knowledge equality, language adoption, and ownership.
This works in any type of social group. If you want to assimilate newcomers you want to:
- be sure they mix well with existing members.
- reduce the skill/knowledge gap between newcomers and existing members.
- ensure newcomers adopt the language (idioms, symbols, unique phrases, or in-jokes) of existing members.
- have newcomers quickly take ownership over parts of the group and proactively lead parts of the group.
We can work on each of the 4 in turn:
1) Intermixing
Do newcomers and existing members genuinely talk to one another? Do they interact as equals or is there a superior/inferior dynamic?
This applies to every type of social group. Newcomers are often treated as inferior and lack existing friends to make connections via friends of friends. They might participate once, feel bad, and not return.
For example, the newcomer posts in the welcome discussion/introduction, the existing members either don’t respond or respond in a manner that establishes their superiority (and makes the newcomer feel dumb e.g. “read the manual!” or “this has been posted before!!”). The newcomer never returns.
You can look for evidence of this in your community data. Our discourse platform is very good for this.
Here is an example:
Discourse shows the last time a member was seen, the topics views, posts read, time spent reading and date created.
If you scroll down to several weeks ago, you can compare the date created with last seen. If they’re close to each other, you have an intermixing problem.
For example,
In at least 4 of the 7 profiles created 26 days ago, the member came, read a few articles, and never returned.
In the other 3, the member has been absent for the past 2 – 3 weeks. This suggests a big intermixing problem.
By drilling into the public profiles we can see if they’re making a contribution or not. We can also see what type of contributions they’re making. Are they creating discussions asking for help or not?
Only 4 of the 7 above made an active contribution in the topic. This is the topic we guide members towards– none posted in any of the other discussions. While we’re very successful at getting people to participate once (and responding to their contributions), we lose people shortly after.
The thread is used almost exclusively by newcomers. Existing members don’t jump in and add any of their views.
This is where we need to make a structural change.
Members are interacting with us, not with anyone else. They don’t intermix. There are some basic tactics that work, i.e. tagging in people you want to reply.
However the more we use this tactic, the less effective it becomes. It doesn’t scale well.
There are two possible solutions here:
- Get existing members to participate here. It’s possible, but unlikely. There is no motivation to post here. They have already mentally positioned this as a place where only newcomers post. Trying to persuade people to do something they have decided not to do isn’t usually a good idea (trying to persuade people to do something they haven’t considered is another game entirely). Even if we did succeed, so many separate strands within a single discussion topic would be unmanageable.
- Get the newcomers to do something different. We could try a different discussion topic. But I suspect we would run into the same challenges. We need the first contributions from newcomers to be significantly different. For example, instead of asking newcomers to post in a single discussion existing members have become used to ignoring, we guide them to create their own discussion. We can ask them to highlight their biggest challenge and the kind of help they’re looking for. Then existing members could quickly scan the discussion and see if they could contribute to it.
This is a much bigger ask that posting in an existing discussion. This means the number of newcomers who make a contribution will be far lower. However if they intermix with existing members, they form bonds and are more likely to return and participate again. Overall I believe the number of retained members will be higher. This also means we can focus our efforts on members that have crossed the boundary (initiated a discussion) over those who happened to drift in.
This requires a scalable, structural, change in the following:
- Update the automated notification. Every member receives an automated notification on the site from our community manager. This guides them to participate in that discussion. We can change this to asking them to create their discussion.
- Update the auto-responder e-mails. Every member’s e-mail is added to our mailchimp list (you can do this manually if you can’t automate it). Once on the list they receive 4 e-mails over 4 weeks. It looks like this:
We can change the content in the first of these e-mails to guide people very early on to create their own discussion.
- Update personal nudges. As the community is still relatively small, we can still drop a personal note to each newcomer in the community. Instead of prompting them to take an action, we can ask them questions about their communities, what they’re struggling with, and then suggest they should make a post in the community about it. Having permission from us will probably have an impact in how many take the action.
Dealing with the fear of creating a discussion
However asking newcomers to initiate a discussion is a big ask. They need to have the time and motivation to do it. Time is a case of priorities. If we feel it helps us achieve our goals, then it helps. Asking for help to overcome our single biggest challenge is a good idea, it might help the motivation side. It’s also more motivating for existing members to help the newcomer (‘I can solve his biggest challenge’).
Yet we can’t ignore the fear factor. It’s socially awkward to be vulnerable in a group of people you don’t yet know. We can reduce this fear by creating and showcasing examples of others doing it. We can include these in the messages we send out to members.
An automated e-mail/nudge here might be along the lines of:
Hi {name}
Welcome to FeverBee.
To get started, we would love you to tell us the single biggest challenge we can help you overcome [motivation]. We have some of the top experts in the field in the community who love a good challenge [reduce fear/motivation].
Some previous examples have included:
- Removing trolls in a cost-efficient way (link).
- Selecting a new platform for my growing community (link).
- Keeping members participating (link)
If you can post here, we’ll rally the troops [non-serious metaphor] to tackle your challenge.
Welcome to FeverBee, I’m truly glad to have you.
CommunityManagerName
Aside, another possible approach would be to treat the newcomers like experts. Ask them to introduce themselves and their expertise and invite questions from existing members on that topic.
Fixing intermixing at a structural level
This addresses the core issues at a deep structural level within the community. It scales well too. The more members that do it, the more members that will do it, and the more there will be around to help others in the future.
Even better, the notion of asking questions that help you brings newcomers close into line with existing behavior. Asking is the bedrock of most successful group interactions. Thus you quickly bring the newcomers into behavioral alignment with existing members.
In summary, we want newcomers to participate in a very similar manner to current members to create cultural alignment and reduce a sense of fraction/inferiority.
2) Skill/Knowledge Equality.
The larger the group, the greater the variance in skill and ability.
Every group has a mean (or average) skill/knowledge level for members. Sometimes the deviation from the mean will be low i.e. most people will be around the same skill level. In practice, this is rare.
Instead, we usually see an inverted bell curve with a high distribution of people that are new to the topic/sector at one side, a high number of experts on the other and a dip in the middle.
This is a problem. The greater the deviation from the mean, the more one group (usually the experts) will feel alienated.
The existing members become frustrated by the same, repetitive, beginner-level questions and leave. If they stop participating, there’s not much reason for the newcomers to stick around neither. Remember that most newcomers will be new to the topic as well.
Social groups need an established domain of knowledge as the basis for invigorating, progressive discussions. Broadly speaking, the lower the deviation from the mean, the greater the quality of discussions (and sense of connection).
Is there huge skill/knowledge inequality?
First, you need to get a sense of whether this is unequal.
One method is simply to ask the experts. You should be able to find and ask your experts what kind of topics they would like to discuss. You can check if those discussions are taking place.
A survey that only goes out to veteran members is useful here (you can send out another just to newcomers too – surveymonkey is a simple tool for this).
Another method is to see what kinds of discussions the top members do participate in. You might find they entirely avoid specific topics.
List members by number of contributions and begin looking at where they participate. Is there a difference between what the experts do and what the newcomers do?
A final method is to look at the discussions newcomers initiate in your community compared with those the existing veteran members participate in.
Below, for example, we see a big emphasis on community platforms.
Yet most established members are probably signed in to long-term contracts.
While they might get personal value from helping others (and some of these discussions are active) they don’t get a sense they’re interacting with people at the same level as themselves. They also don’t get much tangible value (or new information). This limits how much time they will spend in the community (and the sense of connection to one another).
Drilling deeper, you would also notice that the very biggest names in our community don’t participate in these discussions. This is a big clue that the community isn’t valuable for them at the moment and we’re suffering from skill/knowledge inequality.
Your challenge is to raise the skill and knowledge level of newcomers as rapidly as you can. You have to do this without dumping so much information on them that they balk and walk.
Overcoming ‘Balk and walk’
Balk and walk is a big challenge in any education course. Most people simply aren’t motivated to go through a comprehensive course to increase their skill level. You can’t make them motivated. You can only wait for the time when they are motivated to learn and take advantage of it.
Knowledge in any sector isn’t a single continuum. It comprises of many continuums. You might be an expert on community platforms but a novice on psychology (if so, come to SPRINT).
Your newcomers will be at different levels along different continuums. You can’t focus your efforts on a single one.
So let’s make a few structural changes
First, we don’t want newcomer discussions to dominate the community. Yet we also don’t want to ban people from getting basic-level help.
This gives us a few options:
- Create a specific place for people to ask these questions. We could have a separate category for technology, for example.
- We can have a designated platform expert people can approach to answer these questions. This shifts the burden to the few instead of the many.
- We can condense basic platform questions into a single ‘Get Platform Help’ thread where the above expert (or anyone else) could participate. The other benefit here is newcomers are likely to read previous responses if it’s in the same thread than if it’s in the same category.
- We can create a shared document/wiki and add each answer from discussions into a live, editable, platform guide and nudge newcomers with platform questions to download it.
All four would probably work well. The last would be the most effective over the long-term. That might mean an editable Google doc, volunteers who continuously update a live document, or some form of wiki.
Newcomers will be motivated to read the doc to get the answers they need. While reading it, they also realize and see a lot of other platform expertise there that continually accumulates. This progresses them rapidly up this knowledge continuum.
You can do this with any topic newcomers frequently ask about until you have a comprehensive collection of useful guides. Any newcomer can read these guides at the very moment they have the motivation too (i.e. before they ask the question)
The Knowledge Transfer Problem
Now we need to tackle the knowledge transfer problem. How do we get the wisdom from existing members to newcomers without this feeling like a tedious chore (i.e. the same repetitive questions coming up)?
We could, for example, create an Ask The Expert series featuring a different veteran each week. Newcomers can ask questions and get help. Make sure you seed several questions and responses early on to set the tone. If your community has only a few hundred members, you might want to do this just once a month.
Create a Google form and send it to veterans asking:
- What is their unique expertise?
- What specific advice can they share?
- What kind of questions would they love to answer?
Only allow your veterans to be experts (flattery is important at this stage). This will require a lot of promotion to get going. After each session, summarize the key actions into bullet points. Give each person that participates a designated expert badge.
This scales well and competition for expert status might encourage more people to put themselves forward as it grows.
3) Language adoption
Language adoption is the measure of whether newcomers speak like existing members. This is important for both identification and acceptance. Existing members will identify newcomers as one of them if they embrace the same language and it’s clear newcomers have accepted the community identity if they embrace the language.
By language, we mean;
Do they talk about the same topics, use the same references, invoke the same phrases and embrace the in-jokes of existing members?
If you haven’t yet developed a strong community culture, this will be difficult to achieve.
Language adoption typically occurs through osmosis. People are exposed to unique elements of the language, ‘get it’, and begin to use it themselves. The exposure is important. Newcomers have to be exposed to it without being forced to accept it.
- Highlight discussions where in-jokes/references are used. Most communities will have universal discussions which everyone can participate in. These often contain plenty of in-jokes. You can embed these at the top of pages and ensure newcomers are more likely to see them.
- Create a glossary of key terms and in-jokes that newcomers can find. I like this example:
- Create a specific guide outlining the community culture. After a member has made their first few contributions, you can send them a guide that outlines the community brand. Outline what personality the community has. Serious? Fun? Whacky? Laid back? Committed? etc.
4) Ownership
Ownership is the ability of newcomers to quickly take ownership over areas of the community. A comparable would be the time it takes immigrants to become homeowners. This is important. It shows newcomers are making a real investment and feel accepted in and by the group.
This shows itself in different ways. How quickly do newcomers proactively put themselves forward for important roles? Do they initiate discussions? Do they suggest ideas for the group? Do they try to lead different areas of the group?
Encouraging ownership
The first step is to create the official (or unofficial) roles.
Every community or social group affords members official or unofficial high-status positions. Sometimes this is just the ‘leader’. Hopefully it’s a collection of diverse high-status roles where people can apply their existing skills, knowledge, and resources to the benefit of the group.
Too frequently, these roles either don’t exist in the group or established members of the group take them. There are many reasons why the latter can happen. Often the newcomers are too shy to put themselves forward. Even when they do, they don’t have the insider knowledge to stand out. More likely the person who determines who gets the roles simply knows the established members better.
If newcomers aren’t putting themselves forward to take ownership over different areas of the group, they haven’t properly assimilated yet.
Designating experts (as per above) is useful. Inviting people to create content or moderate parts of the site is also valuable. The best approach is to identify what newcomers thinks makes them unique (at an individual level) and find ways they can run their own elements of the site based upon that. Having a resident platform expert is a useful designation (that should come with the ability to create content, moderate related discussions, interview experts etc…)
If you’re creating roles either list them on the site so everyone can apply (only 1 or 2 at a time) or headhunt people. This means you scan through profiles of members paying attention to their early comments on the site to see what sort of role they could have within the community.
Fred Wilson writes people are far more likely to remain in a social group if they know how they can contribute to it. Reading this discussion, it’s clear we have some genuine experts here in different fields.
Or this one from Sam (sorry Sam!)
Clearly it would make sense to drop both of them a line and see if they wouldn’t mind being more involved in any security-related questions on the site. Maybe we could give them a community security-expert status which might encourage them to post more about security issues on the site? These should benefit other members, and them to feel valued.
At the structural level then, we might ensure all volunteers, the community manager, and other members know that if they feel any member has made a really good contribution – they can nominate them for expert status. This expert status provides them the ability to create long-form columns, undertake interviews, and otherwise be more involved.
This ‘ensuring everyone knows’ can come in many forms. We could make newcomers aware of it when they join the community in the auto-responders (yet the open rates of these will vary between 20% to 50%). We could include it in direct outreach. We could periodically ask existing members to nominate others for expert status.
Or better, we can do a big announcement each time someone (or a few people) does achieve expert status. This creates a powerful loop effect. More people see it. More people nominate. More people become experts. More people want to become designated experts etc…
Most importantly, this doesn’t rely on the length of time someone has been in a community. It relies purely on the merit and value that they bring. It relies on existing members noticing when someone has made an especially great contribution and nominating them as an expert (can use e-mail or a button for this).
Tackle the cultural assimilation problem at a structural level
Retention is probably the biggest challenge you face. You can make technological tweaks that might help, but ultimately you need to tackle this problem at the structural level. That means drilling very deep into what’s happening in your community and making changes that scale.
You absolutely can’t do this with any fun, quirky, gimmick. Neither will being nicer to your members. You have to drill deep at the structural and behavioral level to make real changes. This is the only work great community professionals can do.
Research the newcomers from three months ago. Work through each of the 4 cultural assimilation elements in turn; intermixing, skill/knowledge inequality, language adoption, and ownership.
You will find cultural assimilation is the key to long-term retention in any social group. Most importantly, you have far more influence over cultural adoption than you might imagine.
