Twitter: Removing Share Counts to Increase Share Price?

Twitter Share Counts
Posted on: October 16th, 2015 by Patrick Hathaway in Social

Twitter announced recently that they are going to close down the API that provides share counts for webpages.

This means that share buttons (like the one above) will no longer display counts, and tools (like URL Profiler) will no longer be able to report on Twitter share counts.

Twitter Tweet Buttons

Out with the old and in with the new.

It’s fair to say that a lot of people are pissed off about it. You can read them ranting in the comments on the developer blog or on this (rather angry) reaction post.

Twitter published a post on their developer blog that explained why they have decided to do this. In amongst some technical and financial considerations, they claim that:

The Tweet Button

Effectively, they are saying that Twitter share count is a shitty metric.

Are Social Share Counts Bad Metrics?

So, according to Twitter, their share count relates only to tweets or retweets of the exact URL, and all tweets are equal.

Let’s face it, this does make the data incredibly gameable.

Fiverr Retweets

In a more legitimate manner, it is common practice to tweet out the same URL/message multiple times, in order to hit users in various different timezones.

You can also use tools like IFTTT to automate tweets if you have messages you wish to share regularly.

Twitter Automation with IFTTT

Every time 2pm on a Tuesday rolls around, score one more for me on the tweet counter.

This sort of practice is pretty common, and definitely shows that tweet count is not a perfect metric by any means.. But we all sort of knew that anyway, right?

I think it is less common knowledge how the other big players calculate their share counts.


In an effort to seem less shit, it seems that Google’s flailing social network deliberately pumps up their numbers by making any interaction count on their counters.

The ‘plus 1’ count is a rather sketchy combination: the number of times the URL has been had the +1 button clicked, plus the number of times the URL has been shared/reshared on Google+, plus any +1 of those shares or comments.

Here’s an example – our update post when we launched Search Analytics integration.

Search Analytics Integration

See the 65 +1s? Most of those came all at once, after John Mueller shared the post on Google+

John Mueller Share

Call me ungrateful, but those don’t really feel like individual +1s. They certainly don’t represent 65 different people clicking the +1 button, which is sort of what you feel like its meant to be counting.


As the pre-eminent powerhouse of social media, you would have thought Facebook wouldn’t need to pull any fast ones here (whereas Google+, well…).

Yet, here’s how they define their share count in the developer area:

Facebook Share Count

The third point is the most cryptic. So someone posts an update (story) on Facebook, which includes your URL, and any likes or comments will get counted in their ‘like’ count.

Seems a bit wishy washy to me.

And what happens when we are all yaying, wowing and angrying?

Facebook Emoji Buttons

7 options, and yet not one to communicate: ‘this is a load of shit’

Are we really going to have share counts for all of them?

Apples to Oranges

The problem with sharing buttons as they are usually laid out is that the implicit assumption is for the counts to be equivalent.

Social Share Counts

As we have seen, this is not the case. So maybe we can accept the point that these metrics are not particularly good at actually measuring anything.

But does that make them useless?

People Care About Share Counts (Because…)

I don’t think they are useless – simply because people care about them.

Because they help you separate the wheat from the chaff

However shitty a metric share count is, at least it gives you an indication whether people like your stuff or not.

We know a bit about this one, since content auditing is one of the primary use-cases of URL Profiler, and social shares are a pretty standard metric that our users take into account. Social shares give you an idea of which web pages have gained social traction, and (often more importantly), which have not.

Separating the wheat from the chaff

Knowing which pages on your website have received no social interaction can help you decide which content to cull, while auditing. Similarly, pages with a lot of shares indicate that you’ve hit upon a topic that has resonated with your audience. While looking for inspiration on content creation, share counts are a straightforward indicator of successful content.

If Twitter really cared about their users having access to accurate data, they’d build “Twitter Webmaster Tools”, where you could associate your account with a domain and then access page level insights on shares and engagement.

Because you can point to them and say, “I did that”

Share counts are often dubbed ‘vanity metrics’, as if they have no intrinsic value, neatly missing the point that ‘vanity’ is a form of self perception. Behind every piece of content shared on the web is an author or creator, who, like everyone, has motivational needs.

Studies into motivation show that one of the strongest motives is purpose – we want to do things that matter, that are significant – that make a contribution.

High share counts can make you feel like you’ve made a contribution, and because they are visible, others can see that you’ve made a contribution, which also plays into your sense of purpose.

Beyond the squishy feeling of wellbeing, this could also hit some bloggers in the pocket. Bloggers that work with brands to help promote them to the masses can no longer point to their share counts and say ‘Look, I iz well good at Twitter.’

Similarly, marketing teams looking to evidence their success may have to look into other ways to secure executive buy-in on their campaigns.

Because they encourage sharing

Studies into why people share show that the reason that people share content has a lot to do with self-fulfilment. Among other things, ‘sharers’ feel that the content they share helps to define themselves to others, and they want to share information that others will find entertaining or valuable.


Share counts can also provide social proof, which helps readers judge the ‘value’ of content and whether it is worth them sharing it (heaven forbid you share something that others think is ‘only ok’).

Of course, this type of thinking can also lead to ‘blind tweeting’, where readers share articles without actually even reading them.

A Slap In The Face

So, yes, tweet count is a metric of dubious quality, but people care about it anyway. The fact that Twitter know this and are removing it anyway feels like a bit of slap in the face for the community which supported its growth.

Lest we not forget that many of Twitter’s core interactions – such as hashtags, retweets and @mentions – were originally developed by users of the platform, and then later formally adopted by Twitter and built into the UI.

Similarly, before Twitter officially launched their Tweet button in August 2010, everybody used the 3rd party ‘retweet’ button by Tweetmeme (now defunct), which displayed its own counter. 

Now I Ain’t Sayin’ She a Gold Digger…

Some commentators have argued that Twitter’s motive is to encourage more users to login to the Twitter app and access Twitter Analytics, where they can attempt to sell ads to you.

Whilst I’m sure many users would actually welcome share data per URL in Twitter Analytics, there is nothing even close to that in the platform right now. The data is instead segregated per tweet, follower, ad, card, video, etc…

Twitter Analytics DashboardAccording to Twitter, the depreciation of the Twitter Count API actually marks the culmination of a much deeper plan for their API and infrastructure. Twitter are finally completing the process of migrating away from their current database (‘Cassandra’) to their new, real-time, distributed database (‘Manhattan’).

The tweet count feature exists on Cassandra, and so effectively their decision is to not rebuild it on Manhattan.

But if you are moving over to a new infrastructure, why would you deliberately make it worse?

But She Ain’t Messin’ With No Broke…

The only logical answer is that they don’t want to pay to support it. The bandwidth of every blog on the internet requesting data from them all the time must add up to a pretty significant burden – for a feature that has no potential to ever make them money.

In the past they have demonstrated their intent with the API to make it more business-focused, which looks set to continue now. Here is an extract (notations mine) from their 2012 announcement post of version 1.1 of their API: Twitter Ecosystem Take a guess where ‘Share Counts’ fits on that graph? You guessed it – top right.

They don’t want to encourage that stuff because it’s not where the money is. The money is in business applications, not bloggers.

Instead of a free-to-all unsupported API endpoint, you can buy the data from Gnip, the social platform that Twitter acquired last year. We reached out to them to find out if we could get the data we wanted, but they don’t have it either (remember, Twitter decided not to rebuild it) – the most they can give you is 30 days worth. For $3000-$5000 a month, paid annually.

So, yeah, the money is in business applications.

It looks like this is just the tip of the iceberg, however, as Twitter are trimming the fat from other areas of their business too, having just unceremoniously culled 8% of their workforce.

Twitter have clearly had to make some hard decisions, because ultimately, as a public company, they are accountable to their shareholders.

Unfortunately for us, it doesn’t matter how useful we might think share counts are, as Twitter is focused on an entirely different metric – their share price.

Patrick Hathaway

By Patrick Hathaway

I seem to be the one that writes all the blog posts, so I am going to unofficially name myself 'Editor'. In fact, I think I prefer Editor-in-chief. You can follow me on Twitter or 'encircle me' on .

If You Like the Sound of URL Profiler,
Download a Free Trial Today

(You'll be amazed by how much time it saves you, every day!)

  • Free 14 day trial (full feature)
  • No credit card required
  • License from only £12.95 a month


  • I talk a bit about the reason in the Builder Society thread, but the main technical reason is they are moving away from Cassandra completely and on to their own database named “Manhattan”. The count was the last thing on Cassandra. They also stated they “could” recreate it on the new database but it would take away resources (developers) that are working on “more important” things for twitter.

    So they pretty much don’t feel like re-creating the functionality and think their time can be used else where. I follow twitter closely since they are the source of most of our backend database architect and their database people have helped me in tremendous ways with scaling.

    I personally don’t like the move and think it’s a misstep on their part. Their proposed “reasoning” now group themselves in with regular other social platforms which does not showcase what made them unique. There goes their unique selling proposition. Meh.

    • HathawayP

      Yeah, that was pretty much my conclusion as well – they could spend the money/resources recreating the count, but they don’t want to.

      It would be kinda hilarious if social proof is way more powerful than anyone actually thought, and removing the count leads to massively lower tweet numbers, less active users, less ad revenue… and they had to go and add it back in.

      I think you are right about their USP – all the changes they are looking to make could potentially erode it completely.

  • 30-days of Twitter data at that price? I don’t see the value, but perhaps that’s just me. Personally, I always feel share counts are a vanity badge, but I’m first to admit I check regularly after posting anything that requires effort.

    Great post BTW

    • HathawayP

      Thanks Shaun. Yeah – case in point I found myself coming back to this post a couple hours after I published, just as a ‘quick and dirty’ check to see if anyone had shared it. I don’t think anyone is really pretending the metric is perfect, but it is useful.

      To be fair to Gnip, they are offering a lot more depth with the data, but for how we use it normally, it’s just overkill.

  • Tom Binga

    Social counts is a weak metric. The number of leads and revenue trumps everything, and is all a marketer or your boss cares about.

    At the end of the day, if somebody like your stuff and they are prone to sharing things on social media, then they are going to share it regardless of how many shares are showing in your share widget on your website or blog.

    I’m calling bull on the fact that this is actually a big deal. Just my opinion.

    • HathawayP

      You are right it is a weak metric, and not even close to leads or revenue. However the point of a lot of social targeted content (e.g. blog posts) is not necessarily to deliver leads or revenue, at least not directly. So a measure of ‘social success’ is useful, at the very least.

      I am personally not really a big fan of ‘social proof’ actually driving shares – I feel like you that people ‘should’ be sharing stuff because they appreciate it. But there have been numerous correlation studies showing that it does have an effect.

      You are probably also right that it’s not a big deal. In a few months, maybe everyone will have moved on from this or found some other metric to use. As a marketer I think it is a shame since it is a loss of both data and transparency, both of which are good things.

  • I think it’ll depress Twitter’s stock price and here’s how: Like share counts or not, there is a tons of research indicating people share stuff that’s already been shared. It’s social proof, as you reference. You remove the share counts and less people will be sharing content on Twitter. Maybe that’ll bring back “the conversation” Twitter lost, which I bet is at the heart of the purpose here. That might have worked in 2012, but it’s not going to work today. We’re way beyond that. Prediction: within three months, Twitter will bring share counts back.

  • This is very interesting.

    This article is the first place I’ve read about this, so, obviously the first reaction is WTF!

    Now it’s just interesting. Here’s how. No matter how confident the Twitter folks may seem, this is an experiment with no results that could be confidently predicted by anyone. It might be that those little numbers will mess everything up for them and people will click on Facebook buttons instead because it has 500 next to it. Or the opposite – who knows. Twitter is a huge company and they move much slower than in their early days, but it’s still a Silicon Valley startup, so if they know what’s good for them they’ll be agile and able to steer with the wind. So let’s see.

    Another thing that’ interesting trend is the disappearing comments: http://www.wired.com/2015/10/brief-history-of-the-demise-of-the-comments-timeline/

    So, in which direction are we going now, in terms of user participation? It almost feels like someone is trying to make websites obsolete and have everyone just chat on social media and shop on Amazon. Hopefully things don’t end up this way.

  • Patrick hi,
    Thanks for this interesting article which nicely brings together pros and cons of Twitter’s URL share count.

    Might I suggest that within Twitter analytics the data on ‘influencers’ for shared URL is interesting in some industry spaces? Presumably this is also going away.

    What is the data source for ‘why people share’?

    Immediately below the Twitter analytics screenshot you mention the ‘search’ API. I presume this is a typo and should say ‘share’?


    • HathawayP

      Hi Jane,
      Thanks for the typo spot – fixed! The ‘Why People Share’ data came from Chris Dyson’s blog – he is a fellow SEO and an extremely smart one at that (I used to work with him). This was the post I linked to – http://tripleseo.com/why-do-people-share/

      I doubt any of the data within Twitter Analytics is going away – as that is closely tied to their Ads platform and probably a good ‘segway’ for them to onboard users into buying ads with them. It is also unlikely the developed that feature on the old architecture which they are now replacing.


Ready to take your content auditing seriously?