Employee advocacy, and why most programmes stall

A company page with eleven thousand followers reaches a few hundred people. Forty employees with four hundred connections each reach a different few hundred each, from accounts their audience actually trusts. That arithmetic is why employee advocacy exists, and it is also why so many programmes are launched on it and then quietly abandoned.

This is the complete guide to employee advocacy: what it is, why the maths is real and the execution is hard, what actually gets people posting, how to measure it without asking anyone to install anything, and the four ways a programme dies.

TL;DR

What employee advocacy actually means

Employee advocacy is the practice of having the people who work at a company publish under their own names about the things the company does. That is all it is. The definition is worth stating flatly because the category has been colonised by software that redefines it as a content distribution queue, which is a much smaller and much less effective thing.

The reason it works is not reach arithmetic, although the arithmetic is favourable. It is that a post from a named person with a face and a job history is read differently from the same words on a company page. Readers extend trust to people and withhold it from logos, and no amount of budget moves that.

It follows that anything which makes an employee post look like a company post destroys the mechanism it was built to exploit. This is the single most common failure and it is built into most of the tooling.

The moment forty people publish the same paragraph on the same morning, the audience learns to skip all forty. The programme is then worse than nothing.

The arithmetic, done honestly

The pitch for advocacy is usually a multiplication: forty employees times four hundred connections is sixteen thousand people, against a company page of eleven thousand. It is a real effect and it is routinely overstated by a factor of five, because four things shrink it, and every one of them is left out of the slide.

• Overlap. Colleagues are connected to each other and to the same industry. The union of forty networks is far smaller than the sum of forty networks.
• Participation. In an unmanaged programme, somewhere between five and fifteen percent of the people invited post more than twice. Plan for the real number, not the headcount.
• Distribution. Reach is not evenly spread across a network. A post reaches the fraction of connections the feed decides to show it to, which is usually a small one.
• Decay. Participation is highest in week one and falls for eight weeks. A programme measured at week three reports a number it will never see again.

Multiply the headline number by participation, then discount for overlap, and a forty person company with a serious programme is realistically reaching a few thousand additional relevant people a month. That is still an excellent return for the money involved, which is close to none. It is simply not the number in the slide.

What actually gets people posting

The instinct is to reduce friction: give people the words, give them a button, make it take four seconds. That instinct is wrong, and it is wrong in an instructive way. Friction is not the binding constraint. Fear is, and the second constraint is that nothing visible happens when somebody does take part.

Most people do not post because they do not know what is safe to say, they do not want to sound like a brochure to people who know them, and they suspect nobody will react. A one click share solves none of those. It solves the problem they did not have.

1. Say what is safe. A one page list of what people may talk about freely removes the largest blocker in an afternoon.
2. Give raw material, not paragraphs. A fact, a customer sentence, a number. People write their own version and it sounds like them.
3. Make the first post easy to survive. Comment on somebody else's post before publishing anything. Nobody has ever regretted a comment.
4. Have somebody react within the hour. The first post landing in silence is the most reliable way to end a programme for one person permanently.
5. Never mandate. A required post reads as required, and the audience is very good at spotting it.

The full playbook for the getting started part, written for the person who has to actually run it, is in how to get employees to post on LinkedIn.

Measuring it without installing anything

Measurement is where most programmes get stuck, for a structural reason. The accounts you want to measure are personal profiles, and the people who own them will not connect a work tool to them. That refusal is reasonable and it is not going to change.

So the practical constraint is that advocacy has to be measured from the public side of a profile, exactly like a competitor. That is a smaller data set than the network gives an account owner, and it turns out to be enough.

The top row is the whole programme. Participation is the metric, everything else is commentary, and the week participation falls the programme is in trouble regardless of what the engagement chart says. The method, including the arithmetic, is in how to measure employee advocacy.

There is one measurement rule worth writing on the wall before the first review. Never compare two people to each other. An account with four hundred connections and an account with nine thousand are playing different games, and a leaderboard that ranks them teaches the smaller account that taking part is pointless. Compare each person to themselves last month, and compare the programme to itself last quarter.

The second rule is to record what was promised alongside what happened. A team that published eleven posts looks fine until you know it intended to publish twenty, and a team that published six looks weak until you know six was the plan. Without the promise column the review becomes an argument about whether the number is good, which nobody can win.

Whether you need software at all

The advocacy software category is built around a specific promise: a place where employees receive posts and publish them. Whether that is worth buying depends almost entirely on which of the models in the table at the top you are running.

If the programme is a distribution queue, the software is the programme and you should buy it. If the programme is people writing their own posts, the queue is unused within a month and what you are left paying for is a leaderboard.

• You are pushing content to a large, mostly non marketing workforce. The queue is genuinely useful, and licence per seat is the honest cost model.
• You have twenty people who write their own posts. You need to see who published, not a publishing tool. Those are different products.
• Nobody has installed anything yet. Start without software. Everything above works with a shared document and a weekly look.
• The programme already stalled once. Software will not restart it. A named owner and a reason to post will.

The shortlist, sorted by what each product can actually see, is in top employee advocacy software, and the case for running one with no app at all is in employee advocacy without an app.

What good looks like at three sizes

Advocacy is described as one practice and behaves like three, because the binding constraint changes with headcount. A programme designed for the size above yours will fail for reasons that have nothing to do with your people.

The useful question at every size is the same. Who is the one person who will notice on a Monday morning that somebody has stopped posting, and what are they allowed to do about it?

At the smallest size the founder is usually the programme, and the risk is that it stays that way. When one person accounts for most of the output, the company has a personal brand rather than an advocacy programme, and the difference shows the week that person is on holiday. The wider version of that problem is in founder led marketing.

At the largest size the failure is uniformity rather than silence. Enough people post, and they all post the same thing, because a communications team has quietly become the author of forty accounts. That is the template collapse described below, and it is much harder to reverse than a participation problem.

What it is not: influence, social selling, and thought leadership

Three neighbouring practices get bundled into the same budget line and they have different mechanics, different owners and different success measures. Confusing them produces a programme that is judged against the wrong number.

• Employee advocacy is breadth. Many people, occasional posts, the goal is that the company is spoken about by people rather than by a logo.
• Social selling is depth. A few people, a specific audience, and the goal is a conversation with a named buyer rather than an impression.
• Thought leadership is authority. One or two people, sustained over years, and the measure is whether the industry quotes them.
• Influencer work is rented audience. It is a media buy, and it is the only one of the four where the account does not belong to you.

The practical consequence is that an advocacy programme should never be judged on pipeline in its first two quarters, because breadth pays back slowly and the fastest way to kill one is to ask it for a number it does not produce yet. The version of this practice aimed at a named buyer is in social selling.

The four ways a programme dies

Advocacy programmes rarely fail loudly. They go quiet, and the quiet is usually explained afterwards as a lack of buy in, which is a description of the symptom rather than the cause.

1. The launch spike. Thirty people post in week one because it is new. By week six it is four people, and nobody noticed the slope because nobody was reading it weekly.
2. The template collapse. Suggested posts get published verbatim, the feed fills with identical paragraphs, and engagement falls for everyone including the people writing their own.
3. The owner leaves. Every working programme has one person who reads the tracker on a Monday. When that person changes role, the programme ends within two months.
4. The measurement gap. Nobody can say whether it is working, so at the next budget conversation it cannot be defended, and it is not defended.

Three of the four are visible weeks in advance in a single column: posts published per person per week. That is the argument for tracking participation even when the programme feels healthy, because the feeling arrives about a month after the data.

There is a fifth way, and it is the one nobody writes down because it looks like success. The programme works, a handful of people become genuinely good at it, and the company starts treating their accounts as company channels: approving drafts, asking for a post before a launch, correcting tone. Every one of those requests is reasonable on its own, and together they turn a person back into a page. The people it happens to usually stop posting within a year, and they leave with their audience.

The defence is unglamorous. Write down what the company may ask for and what it may not, agree that the account belongs to the person including when they leave, and accept that a programme built on borrowed trust cannot also be controlled. Companies that refuse that trade get a distribution queue, which is the thing they were trying to avoid.

How Groowth does this

Groowth reads the public side of a profile, which is exactly the constraint an advocacy programme runs into. You paste the public handles of the people taking part, and it records what each of them published and when, their follower count, and the public reactions on each post. Nobody connects an account, nobody installs anything, and nobody has to be asked twice.

What it shows is participation first, because that is the number that predicts the rest. It never shows reach or impressions on a personal profile, because those are given to the profile owner and nothing outside can see them. Three accounts are free with no card, which is enough to watch a pilot before deciding whether the programme is real.

Frequently asked questions