The Corruption of Reciprocity Online
I wrote the first draft of this for a guest lecture to a college class called Psychology and Social Media. Just combining the two topics of reciprocity and social media, was enough to meet the focus of the class and keep the student’s attention.
But then I started to think about the pattern of reciprocity from a product designer’s perspective — what if you worked at Twitter?
There’s a pattern of abuse working under the name growth hacking that feels healthy at first and then becomes destructive to the product. Should you treat growth hackers in the same way that you treat spammers?
And, what do you do if you’re a marketer? The tactics in this article work, or rather worked for awhile. Is it not your job as a marketer to use them?
And then I thought about it from the perspective of the users of social media products. That’s all of us. Reciprocity is yet another way that you’re being manipulated online. I’ll get to a conclusion in the end but I’m also very interested in what conclusion you come to.
Anyway, here goes.
Reciprocity is a social norm of responding to a positive action with another positive action, rewarding kind actions. — Wikipedia
When someone gives you a birthday gift on your birthday and then, later, you give them a birthday gift on their birthday — that’s the good kind of reciprocity. That’s the sort of social norm that holds together a healthy society.
I want to look at the underbelly of this social norm and the way that it repeatedly comes up as a pattern of manipulative influence on social platforms.
For those of you who have read Cialdini’s book Influence, he mentions a key quote about how reciprocity can be used to manipulate people (although the original source is Marcel Mauss).
“There is an obligation to give, an obligation to receive, and an obligation to repay. Although the obligation to repay constitutes the essence of the reciprocity rule, it is the obligation to receive that makes the rule so easy to exploit”
Yes, people have been exploiting reciprocity for a long time. What’s different on social platforms is scale.
In August 2002, a popular maker of blogging software, called SixApart, introduced a new feature called trackbacks.
Trackbacks allowed a blog author to receive notifications when another blog linked to one of their posts. That link was generally perceived as an ego boost or, in the language of reciprocity, a positive action.
Imagine you wrote a blog post that you were very proud of. Then, another person quoted you in their own blog post. That would feel good, right?
So, this invention of trackbacks created visibility into a positive action and that visibility triggered reciprocity among bloggers. If you link to me, I’ll try to find a way to link to you.
If you come from academia, this is basically identical to having someone include a citation to your research (and reciprocal citations are a thing). However, we in the blogging world gave this behavior a more colorful name, Link Love. If you love me, I’ll love you back.
If you didn’t live through this period online, this was a very different social media landscape. There was no Facebook or Twitter. Blogging was how people social-media’d each other. It was, comparatively, a much simpler time.
Link love is the earliest online example of reciprocity at scale that I can think of. Although possibly, email spammers were taking advantage of it earlier than 2002. Regardless, I consider this to be ground zero of the industrialized corruption of this reciprocity.
In the world of blogging, link love wasn’t merely an exchange of ego boosts. A link came with two additional benefits.
Links grew subscribers to your blog
Linking to another blog generally caused a few of your readers to click through to the other person’s blog and, if the blog was good, subscribe.
In the early days of Google, your ranking in the search results was primarily determined by how many people linked to you.
So, if two people wrote a story about the Oscars, the site that everyone linked to would be the one that showed up first in the results. This was the original Google patent that made their search results superior to everyone else. Google called it PageRank.
Both of these benefits just come down to ways in which a blogger can get more readers.
The supposedly generous act of crediting another person’s work through a link created an obligation for them to link back to you. You didn’t lose anything by linking out to other people and, through the power of reciprocity, you could expect to gain readers through triggering other bloggers to link back to you.
So, a naive writer might think: “If I write really well, readers will come.”
But a savvy writer would think something very different. They would think that if they link to as many blogs as possible, many of those blogs will link back. (Because of reciprocity.)
These reciprocal links will each send subscribers and additionally increase the savvy writer’s ranking in Google search. Therefore, savvy writers should adjust all their posts to have as many links to other blogs as possible.
That behavior turns articles from products into marketing tools — the more you write the more your blog grows.
In this era, writers were only lightly affected by this new marketing behavior and still felt an obligation to write good and interesting posts. But their writing had begun to be corrupted by artificially working in links to other people.
The phrase we use now for this corruption is growth hacking. That’s just a peculiar form of marketing where you are manipulating the social systems of internet platforms.
Trajectory of a Growth Hack
The trajectory of a growth hack is that it starts out as a simple observation, then it becomes a trend, and then it becomes an industry.
At the point that the term link love was coined, the reciprocal linking growth hack was merely a trend. It was a commonly known tactic, used regularly. But it didn’t feel overwhelming.
But, shortly after the term was coined, people started building businesses to turn this hack into an industry.
By 2004, there were conferences, agencies, and publications covering what they called search engine optimization (SEO) and search engine marketing (SEM). SEO and SEM were the earliest forms of growth hacking on the web.
There were two key strategies in this SEO/SEM industry, both with the goal of manipulating Google into sending your site more readers. One strategy revolved around keywords — that’s not what I’m talking about.
The other way revolved around getting more links to your site. Link love is one way to get more links.
At first, SEO professionals tried the link love strategy as I originally described it. They’d write a decent article that links to you in the hope that you’d link back to them.
Then, these new SEO-folks started to get impatient and would email you asking if you’d link back to them in exchange for them linking to you. This was no longer reciprocity. The offer had morphed from a gift to a simple transactional business deal.
Then, SEO professionals became really impatient. And, in that moment, morality went out the window. Both strategies above involve the hassle of waiting for another, possibly moral, person to respond to your link with a reciprocal link.
What if you could take that slow-moving moral person out of the equation? So, the SEO industry had a moment where they would just go and spam a blogger’s comment areas with links.
And then, when bloggers tried to shut that down, SEO people became 100% clear that other bloggers were always going to be a huge hassle.
So, they created something called content networks. Why reciprocate with another person’s blog when you can just create two blogs and reciprocate with yourself? A content network can be hundreds or thousands of supposedly unrelated sites that then create a network of links to each other.
So, just to clarify here. Content networks don’t have anything to do with reciprocity unless you’re viewing the networks through the lens of history.
That’s exactly the lens I want you to be using. Reciprocity laid the groundwork for a massive attack on the way the system of search results works.
Result of the Attack on the Search Ecosystem
Here’s the result of that attack on the search ecosystem.
Google’s search was the best because they observed something profound. The number of links to a site is a good proxy for the value of that site’s information.
But then this observation got corrupted, first through reciprocity, then through the industrialization of that reciprocity.
The growth of link building hollowed out the value of Google’s inbound-links-equals-authority observation.
Google’s first reaction to content farms and comment spam was to try to detect and remove them from the Google algorithms.
By 2009, Google had realized that they needed a different algorithm and now they look at many more factors than inbound links.
At the same time, so many bloggers had read so many growth hacking articles that none of them continued to believe that a link was an ego-boosting gift.
The history of link love is that it emerged from reciprocity, was professionalized, industrialized, and then killed the ecosystem it relied on. Bloggers changed their social norms and Google changed their algorithm.
I’ve heard this pattern described by other marketers as a scorched earth approach.
“We cross the line into a sort of scorched earth marketing mentality where we forget the reason consumers were drawn to that channel to begin with. “— Kipp Bodnar in This Is Why We Can’t Have Nice Things.
You know other marketers are going to eventually cross the line, destroying the opportunity, you might as well be the marketer who crosses that line first and fastest.
That pattern repeats itself: reciprocity, professionalized reciprocity, industrialized reciprocity, and then death of the host. That death is so crucial to understand.
The industrialization acts as a cancer. In email, we treat spam as a cancer and thus blackball the people behind spam. They are not welcome in mainstream tech society.
However, what are we doing with growth hackers? We actually celebrate them, even though they exist to pollute, corrupt, and eventually kill our favorite platforms. Link love is just the first example.
The early social networks were called Friendster and MySpace. These are the two big ones before Facebook.
All three social networks were built around a synchronous follow model, which means both people needed to opt-in before they were connected in the network.
I couldn’t just follow a celebrity like Dave Chappelle unless he wanted to follow me too. Usually, this only happens when you are friends or family. So, most social networks just called this feature friending.
It wasn’t altogether too difficult to navigate the social norms of a friend request. If a friend friended you, you said yes. This wasn’t so much reciprocity as it was informational. “Hey, my cousin Billy is on Friendster now.”
Meanwhile, if a stranger friended you, that was creepy. A stranger friending you isn’t doing a positive act, they’re doing a creepy act. So, this didn’t trigger reciprocity.
There’s a grey area where a light acquaintance “friends” you. Many people struggled with this, but not because of reciprocity. Rather, they were annoyed by the request but were afraid of appearing rude.
However, when Twitter launched in 2006, it introduced asynchronous following. That meant that I could follow you without you following me back. You didn’t even need to approve my follow.
Twitter gets you all excited about a new follower because receiving a new follower notice feels like receiving a positive act — someone is interested in you, therefore you must be an interesting person.
Twitter sends you an email and makes it easy for you to learn all about this new follower in case you want to follow them back.
This is an intentional product decision by Twitter. If this reciprocity wasn’t serving them, Twitter could alternatively just tell you your follower count changed. Yesterday it was 10, now it’s 11. That would generate an ego boost without triggering reciprocity.
But instead of doing that, Twitter very specifically creates the reciprocity dynamic by telling you who followed you. And that design decision created a bit of confusion in the beginning, because a person following you feels like it might be the gift of a positive act. Do you need to reciprocate?
Not all people, but some people, felt very strongly that if I followed you, you should follow me back. There was almost a built-in idea of reciprocity. To not follow back would indicate that you’re a rude person.
Of course, now that we know what Twitter is, following some celebrity doesn’t obligate that celebrity to get involved following the details of your life. That would be a crazy expectation.
But in the beginning, people were tripped up. And many people never got the message. If you followed them, they would follow you back even if they had no reason to want to read what you were tweeting about.
In an experiment in 2015, the publication Blog Tips created a nonsense Twitter account that posted only complete and total gibberish.
And then Blog Tips used this gibberish account to follow 400 of Justin Bieber’s most recent followers. How many of those Beliebers would follow back? Blog Tips reported that one-quarter of those people did follow back. That’s the power of reciprocity.
Following Is a Positive act
Following is a positive act. It’s a dopamine hit. On Twitter, following is clearly an act that triggers reciprocity. That means following can be manipulated.
If I follow you, you will follow me back. Then, because you’re following me, I have a shot of getting my messages and links into the activity feed that you read.
And, because there is no notification when I unfollow you, I can just unfollow you in a week or so. That way I don’t have to read your stuff.
This pattern became a trend with a name. It’s called follow-unfollow.
First, you follow a ton of accounts, then you unfollow those accounts. This is the professionalization of reciprocity on Twitter. Follow-unfollow is a named marketing tactic that all marketers are aware of.
And on Twitter (and soon other places), follow-unfollow is following the pattern of observation to trend to industry to corruption.
You can often spot people who have become prominent through Follow-Unfollow because they are not as aggressive about unfollowing as they are about following.
For example, on the blogging platform Medium, the 33rd most followed author is followed by 88k readers. This author follows 112k readers. That looks like someone who used follow-unfollow to achieve prominence, only they were more aggressive about following then they were about unfollowing.
Just like link love helped launch the SEO industry, follow-unfollow also created a new industry. This industry is served by conferences, by publications, by agencies, and by software services generally living under the umbrella term growth hacking.
Theoretically, having a lot of followers gives you credibility. You look like a famous person.
And then, also theoretically, having a lot of followers means you have a lot of readers. And so, you can push your ideas and products to them.
So, the goal of the follow-unfollow industry is to grow your audience.
On Instagram, one playbook is for a marketer to post random quotes on top of images. The posts look something like this:
- It costs $15 on the service Fiverr to get a unique set of 2000 of these quotes and images.
- Then, the marketer will automate the posting of these images using a marketing automation service.
- Then, the marketer will turn on an automated follow-unfollow bot, often using another marketing automation service.
- Over the course of months, this will grow a generic Instagram account into thousands or even millions of followers.
- Then, the marketer will just sell the account to someone else. For $30k, you can buy a ready-made Instagram account with one million followers.
That playbook relies on several services. Many of these marketers are reading about tips and tactics on GrowthHacker.com, they are using an ecosystem of content generation based on services like Fiverr and Upwork, they are using automated software tools for posting, and they are expecting to make money through marketplaces for Instagram accounts, like FameSwap.
That’s what I mean by industrialization. There’s now a professional industry around this follow-unfollow reciprocity tactic complete with businesses offering tools, marketplaces, and serious money changing hands.
Industrialization Hollowed Out the Reciprocity Mechanism
Now, with link love, I made the case that industrialization hollowed out the reciprocity mechanism.
For example, I tested a service called HypeGrowth on Twitter. Over a period of a few months, HypeGrowth followed people who tweet about topics that matter to me with the hope that some of those people would follow me back. It generated 3,000 new followers.
However, I can’t think of a single new follower who has interacted with my tweets. All I see are the same followers that have been tweeting with me for the last five years.
This made me wonder though: is audience building about vanity or performance?
Audience Building — Vanity or Performance
On Twitter, there is one way to estimate how valuable an audience is. It starts with the question: “If I post a link, will people click on it?”
If the answer is yes, then you want to grow your audience as big as possible.
So, I looked at one semi-famous person who has been on Twitter for a long time. His name is Tim O'Reilly. He’s famous to me because I used to work for him. And he’s famous to a certain segment of tech and business people.
But maybe you’ve never heard of him, so I’ll just give you his follower numbers. In 2010, Tim had about 500k followers. (I’m estimating this based on the date that Ashton Kutcher became the first Twitter account with a million followers.) Today, Tim has just about two million followers.
Between 2010 and 2018, Tim posted 1016 tweets that included a link that was shortened by the bit.ly service. Now, you don’t really need to know that much about bit.ly other than that I hacked the service to see how many people were clicking on the links this famous person was tweeting.
(Although, if you have to know, bit.ly is known as a URL-shortening service. You post a long link and bit.ly returns something short, like bit.ly/XYZ123. The hack is to know that you can add a plus sign to that link, making it bit.ly/XYZ123+, to see analytics on how many people clicked the link.
Technically, I wrote code to connect to their API, but that’s just the programmatic way to do something you could do manually, even if you aren’t a programmer.)
Using this bit.ly hack, I can see the average number of people who click on one of this person’s tweets. Here’s what that graph looks like for Tim O’Reilly:
Let me summarize that in plainer language:
- In 2010, 800 people would click on each of his links (on average).
- By 2018, his audience had grown from 500k to 2M, but the performance of his tweets had shrunk to just 400 clicks.
What you’d hope is that, by growing your audience from 500,000 to 2,000,000 followers, you’d also grow the number of people who clicked on a link. The reality, for Tim, has been the opposite. He grew followers, but lost clicks.
There are two obvious explanations and my guess is that are all happening at once:
- Several former Twitter employees told me that the culture of Twitter moved away from clicking on links. People are there to talk to each other, not to read long-form articles. So, maybe that’s the entire explanation.
- While Tim’s follower count grew, so did the follow counts of each of his followers. If each of your followers only follows you, you have 100% of your attention. But each of your followers is also following lots of other accounts, then your share of their attention drops.
In any case, it’s alarming to see how such a big increase in your theoretical audience result in an actual reduction in the value of that audience.
This looks like the industrialization of follow-unfollow has hollowed out the value of the size of your Twitter audience.
Follow-Unfollow Could Be Destroying the Value of Audience Building
I don’t have as much data on Instagram, but I assume the same pattern is happening there because I know the same follow-unfollow tools are used there.
On these social networks, the relative size of your audience is especially valuable if it’s bigger than other people’s audience.
But, since everyone else is growing their audience and everyone in their audience got there because of their propensity to follow out of reciprocity, each of your audiences is made up of very distracted people.
So, that’s a simple way that follow-unfollow could be destroying the value of audience building — it simply creates more noise so that your posts are less likely to be seen.
In the example above of Tim O’Reilly, a follower is now worth 1/8 of what they were worth in 2010. Based on my one, cherry-picked data set, the value of an audience on Twitter dropped 87.5%.
And, just like Google adjusting their search algorithms, Twitter and Instagram are responding by adjusting their feed algorithms.
The adjustment that Twitter and Instagram are making is to move to algorithmic feeds. They will show you posts that are not from people you follow. And they will find posts from people you do follow and then either hide them or promote them based on how much people engage with them.
The history of follow-unfollow is that it emerged from reciprocity, was professionalized, industrialized, and now is being deprecated.
But it’s not dead yet. I recently ran a follow-unfollow test on a social media site that I care about and found that the follow-back rate is 14%. If I follow 100 people, I will grow my audience by 14 people. And then, next week, I can unfollow all of them.
We are entering a world of mass-manipulation because of the scale of these social platforms.
Hopefully, it was worth reading this deep dive on one way you are being manipulated.
I’m also hoping that this will reach some product designers and have them think about the way in which growth hacking destroys their product.
The design of a product’s features and APIs has a big impact on how ripe the product is for growth hacking. So does the way we measure the value of our products — a page view is not necessarily the sign of a satisfied customer.
Right now, we (in the tech industry) treat spammers with an iron fist. We fight like hell to keep them off our platforms and to minimize their negative impact.
And yet, we treat growth hackers as an acceptable type of user. Why is that? You build a product and you do whatever you can to make it a great product and then you’re just going to allow some growth hackers to come in and corrupt it?
If you’re a growth hacker though, what else are you going to do? If your product is better than the competition, it’s your job to get that product in the hands of those customers. I don’t fault the growth hacker.
What would you change about yourself, these products, or the way marketers operate?