media attribution is BIG news. 

Marketers are struggling to attribute revenue to social channels, and lack of definable ROI is one of the major reasons that businesses cut back on social investment. 

I spend a lot of time looking at our own social attribution, but it strikes me that in many cases, the closer I look, the less clear a picture I have.

This isn’t because the figures I have to work with aren’t clear.

It’s because, in a lot of cases, they might not be true...

How much traffic does Twitter provide?

Twitter is our busiest social channel here at Econsultancy. I’m going to run through a few examples in this post, and while they seem to occur across a lot of social channels, Twitter tends to be the channel where they are most obvious, mainly due to the varied, fast and furious sharing of links and the proliferation of apps and third-party clients in use. 

Before we begin, I'm aware that I might be making a massive mistake somewhere here - am I missing something obvious? Let me know (please!). But for now let's assume that I vaguely know what I'm doing. 

To start with, here’s a simple sounding question: How much traffic does Econsultancy get from Twitter? 

The answer to this should be simple enough. We use Google Analytics on the site. So If I go to ‘All Traffic’ and type ‘Twitter’ into the filter box, I should have my answer. 

Here’s what happens if I do that for the period Feb 5th – Feb 10th, 2013:

3,416 visits.

Given that we have 126,000+ Twitter followers, that seems awfully low doesn’t it? 

What about links? 

If you attach any link to a tweet that is longer than 19 characters, Twitter will automatically wrap it in a link. So we need to look at those as well: 


That’s a bit better. But all links don’t come from Twitter. Some are from Pinterest, and some are from Reddit, and some are from other mysterious sources, because occasionally people copy and paste them from Twitter to somewhere else. 

So that figure is actually 9,605. Plus the 3,416 from ‘Twitter’.

So 13,021 total over five days. 

I know that there are other sources as well. Some of the traffic from clients like Hootsuite, Pocket or Instapaper will all be grouped separately (and keep that ‘some’ in mind). 

Does 13,021 still seem low?

Drilling down

Maybe. So let’s look a little closer at how individual tweets perform.

I make a point of retweeting our posts so that our followers in different territories have a chance to read them. I schedule them  at least eight hours apart so hopefully you’ll see minimum duplication if you follow us. 

If I check in Twitter, I’m provided with figures showing how many people clicked on each of our tweets.

For this post, I’ve chosen three random tweets from the 5th of February. Each was tweeted twice:

As you can see, I’m provided with click info next to each tweet. The gold, green and purple figures represent favourites, retweets and replies respectively.

Given that the first tweet was retweeted 25 times, I’d assume a bare minimum of 25 extra clicks, depending on who retweeted, to how many followers, at what time etc etc, but for the moment I’m going to discard that and go purely on the people who clicked on our tweets

So for each we have:

Post 1: 443 Clicks

Post 2: 597 Clicks

Post 3: 446 Clicks. 

Even allowing for the occasional failure to load, I don’t think it would be unreasonable to assume that the vast majority of those people made it to the blog post in question, and would show up in analytics. 

So I go to GA, and select content/all pages/ and drill down using the blog post ID number (you can see it in the url of each post: )

Here’s the first post:

1,230 visits.

And I’m expecting a minimum of 443 to come from Twitter. 

I extracted this as a csv and sorted the traffic sources. 

GA tells me that:

  • 460 came from our email newsletter, the Daily Pulse.
  • 88 came from direct traffic.
  • 442 came from ‘other’ (Organic, Google, PPC etc – other trackable sources). 
  • 240 came from Twitter. 

Which means that somewhere between clicking the link in a tweet and coming to the page, 203 people vanished. 

Surely some mistake? 

Here are the figures for all three tweets:

Every single time, the numbers don't match.

In tweet three’s case, more came from Twitter than clicked on our two tweets, but not that many more. Given the number of mentions and retweets I’ve seen, that seems remarkably low (Just for the record, lists that post as having 111 mentions on Twitter, but that seems like less than average to me). 

Analytics can't handle the truth

This isn’t to say that Google analytics is inaccurate, or that Twitter is giving me false figures.

When we send a tweet it’s reparsed and wrapped in any number of shorteners, passed through clients, through buffers,through dashboards and rss feeds, reposted and then retweeted ad infinitum. It stands to reason that a lot of it would get messed up along the way.

It’s also entirely possible that I’m looking at the wrong figures (Google Analytics experts, I want to hear from you in the comments on this!), but I don’t think I’m doing anything wildly unreasonable.

But that isn’t the real problem here. Let’s look at tweet number two. 

According to our figures, Twitter accounted for 17% of views.

If we add in the missing people it would account for 21%.

And that’s ONLY the missing people who clicked OUR tweets.

I had a quick look at who retweeted it, and they have a combined potential reach of 24,974 people. And as mentioned, that doesn’t include retweets of retweets, or any other six-degrees-of-separation stuff. 

Direct traffic accounts for 18% of visits to that post, and I'm almost certain that some of the traffic from Twitter has become messed up and appeared as Direct.

Those missing Twitter clicks would account for 22% of direct. And Direct traffic currently accounts for 50% of our entire site traffic. 

The attribution problem

Currently a huge number of marketers dealing with social are falling back on woolly metrics because the financial figures they have may not appear to justify the investment they receive.  

You can append tracking code to everything you tweet, but the sharing ecosystem will still muddle your figures. Currently I’m reporting the minimum financial figures for social media each month, and I honestly don’t know what the answer to this is (again, I welcome suggestions on this).

This is first and second-party data. I'd never expect 100% accuracy, but in this case the differences seem unreasonable.  I’ve seen a lot of analytics systems and I haven’t seen any that are significantly better at tracking content once it moves off-site.

The real problems start to appear when we look at attribution. Currently our social channels account for around 20% of our on-site conversions.

Or they might account for much, much more than that. But ‘might’ isn’t going to do me much good when it comes to requesting budget. 

What do you think? It feels as though that social value is being overlooked and denigrated in many cases, and this is reflected in falling investment. Social isn’t the shiny new toy anymore, but it may well be one of the most valuable assets in the marketing toolbox. Do you think there are any genuine solutions here? 

Matt Owen

Published 12 February, 2013 by Matt Owen

Matt Owen is a marketing consultant based in London. He was previously Head of Social at Econsultancy and currently runs Atomise Marketing. Opinions expressed are author's own.

204 more posts from this author

You might be interested in

Comments (49)

Save or Cancel
Nuttakorn Rattanachaisit

Nuttakorn Rattanachaisit, Consultant at Possible Worldwide

If you want to track performance from Twitter - you should use UTM Tagged URL and then use shorten url for sharing over the Twitter. So it allows to consolidate traffic from Twitter campaigns.

over 5 years ago

Matt Owen

Matt Owen, Marketing Consultant at Atomise Marketing

Hi Nuttakorn - yes we do, all our blog posts are appended with tracking code before we tweet them, but they still don't all show up in analytics. As mentioned I'm assuming this is because of the way that links are shared/wrapped/retweeted etc, which corrupts the tracking code as it journies around Twitter.

over 5 years ago

Dmitriy Melihov

Dmitriy Melihov, Project director at SpyBOX

Hi Matt,

we did some internal research early and found that up to 5% of user-agents (or using a proxy) doesn't provide referrer string, so you see all that visits as Direct, they are there.

over 5 years ago

Agustin Mingorance

Agustin Mingorance, Senior Account Director at VelocitySmall Business Multi-user

Hi Matt,

No doubt there are visits falling through the cracks, but there are a few of additional things to consider:

1) We've measured lots of Twitter campaigns and we've never found a strong relationship between activity and visits. It's common to find more retweets than site visits. I think this is a big issue in your case: you've an established blog and quality control that means people will retweet without visiting.

2) Your big list of followers is also worth considering. A large number of retweets will hit people who already follow you. That means you're not necessarily expanding your universe through activity.

3) The numbers clicking through for your campaign might 'seem' low but look way above (as you would expect) standard rates for a hardworking blog post.

4) Your content is attractive to marketers but many will disappear in the blink of an eye if it turns out it's on a subject that doesn't instantly grab their limited attention.

So I'm sure a proportion of your tags are washed away in the sharing process but I'm not sure it's entirely to blame. Worth considering an investigation on two fronts here.


over 5 years ago

Matt Owen

Matt Owen, Marketing Consultant at Atomise Marketing

Hi Neil,

Thanks, yes those are certainly worth looking into. I do think there's a fair bit of 'non-visiting sharing' that occurs, and our blog bounce rate is always fairly high, but I'd expect it to be for most decently trafficked blogs - even so I'd assume they'd at least register as a visit. I may well be wrong, definitely worth a closer look, thanks :)

over 5 years ago

Agustin Mingorance

Agustin Mingorance, Senior Account Director at VelocitySmall Business Multi-user

No worries Matt, really interesting hole to try and fill.

I've often looked at Twitter goals as activity rather than visits and conversions due to the issues you've outlined.

But if we could identify chunks of missing traffic then it would change a lot for me. Great initiative...

over 5 years ago

Ally Manock

Ally Manock, Head of Digital Strategy, Planning & Insight at Brass Agency

Another very good, transparent and honest post - makes interesting reading!

over 5 years ago


john mitchell

Couple of thoughts?

1. Are you a free or premium GA customer? I'd presume premium, but if not what is the sample rate. (im presuming same time periods have been selected.)

2. If all links have UTM source, are you sure they all state twitter when sent to twitter? utm source overwrites the referral, so if a non twitter utm is passed via twitter it will be tracked as the incorrect value.

3.Was there a paywall on any of these links? and is the GA custolmised to track this differently using for instance the trackpageview to overwrite the URL

over 5 years ago

dan barker

dan barker, E-Business Consultant at Dan Barker

I'd said on Twitter, but might as well say here for anyone else struggling with stuff like this:

The econsultancy twitter stream is a mix of owly links & bitly links. Some of those have utm parameters applied, which will overwrite the fact that twitter sent the traffic, some don't.

The blog post also compares page views with visits & various other things that are unlikely to match up too.

That makes it very difficult to tell how much of the hole here is due to the tools themselves, and how much is due to how they're being used. (either on purpose or by default).

If you're struggling with a similar issue, follow these steps:

1. Use 'utm parameters' as Nuttakorn suggested.
2. Put all of your utm tagged links through bitly shortening.
3. Then post them to Twitter.

Having done that, you will then be able to see:

1. The stats within Google Analytics, using the campaign details you set.
2. How many clicks bitly registered.
3. How many clicks twitter registered.

From there, you can compare the 3, along with a breakdown by blog post, by date, by location, etc to be able to triangulate the issue, or at least to be able to get a realistic idea of the scale of the issue.

The data will never, ever match perfectly. As with all data, these tools simply give you an abstract picture of reality, and each is a slightly different abstraction. The useful thing is to understand how much they differ from reality and (ideally) why.


(ps. caveat: the above is looking at it from a measurement angle. looking at it from a performance angle, accuracy often matters far less, as long as it gives you the information you need to improve performance)

over 5 years ago




the "" filter result include also pinteresT.COm, getpockeT.COm... because they contains "" in URL

over 5 years ago

Matt Owen

Matt Owen, Marketing Consultant at Atomise Marketing

Hi John -

Free user. Sample is from 5th-10th Feb in all cases.

UTM tracking may be responsible for some of this, but generally speaking all our tweets go out automatically from the blog, with Twitter as source. Frankly this does need deeper investigation though. I get the impression we need to spring clean some of our analytics reports a bit, but the level of inconsistency does still seem very high to me.

We don't use a paywall so this shouldn't affect anything.

Thanks - all good points and should be considered by anyone tracking Twitter.

over 5 years ago

dan barker

dan barker, E-Business Consultant at Dan Barker

@zagp very good point.

over 5 years ago

Matt Owen

Matt Owen, Marketing Consultant at Atomise Marketing

Thanks Zagp - somebody pointed that out on Twitter earlier as well, at which point I slapped myself in the face - obvious! Worth pointing out though as I'm sure I'm not the only person who does that -I'd always assumed it was because somebody had copied a link and added it somewhere else, but that makes perfect sense.

Dan - excellent, thanks for that, granular measurement ftw!

over 5 years ago


Nick Stamoulis

"It's common to find more retweets than site visits. I think this is a big issue in your case: you've an established blog and quality control that means people will retweet without visiting."

I think Neil has a good point. When you have a powerful brand people will share your stuff without ever actually clicking over to it simply because it's coming from you. I don't know if that would account for all your missing visitors but it could definitely be a contributing factor.

over 5 years ago

James Gadsby Peet

James Gadsby Peet, Senior Digital Services Manager at Cancer Research UK

some of this will also come down to Clicks vs landings on the Twitter figures presumably? we've seen ad campaigns (especially on mobile) that only have a 35% landing rate from clicks. As so much twitter traffic will be from mobile it's likely that this could be causing a bit of discrepancy between the two figures...

over 5 years ago

Matt Owen

Matt Owen, Marketing Consultant at Atomise Marketing

@James - yes definitely, although I'd find it hard to believe it would be to this extent. That said, different clients would be a cause and these are probably more prevalent on mobile, and some mobile OS tend to group traffic into Direct so this would definitely be a factor.

over 5 years ago


Mike Schmutz

One of the problems with Google Analytics is all the recent updating that has changed the backend javascript. One of the things you may want to look at is going into your Social tool and look at the overall landing pages for Twitter. You may find a few of the blog pages that are performing poorly. This could be due to old code still attached to these blog posts that haven't been refreshed since GA updated their tracking. If this is the case the bounces could possibly be occuring and the visits could be off because the tracking code is in the footer and not loading correctly.

Or quite possibly a page on a blog doesn't have tracking code installed and is therefore losing visits to the system. I've seen this happen many times on blogs because of the amount of content being produced. It slips through the cracks.

Overall though systems will never match up, and you'll pull your hair out trying to match system data, it's best practice to choose one type of reporting system for your site and stick with the data it produces, that way you will be able to track overall trends on performance easier.

over 5 years ago


Nick Fawbert

Setting up a different URL for the twitter posts with a 301 redirect?

over 5 years ago



Hello, Matt!

When posting links to Twitter you need to take into account that each click goes through 1) parser and 2) parser and only then results with a website visit. Though both of them are fast, they still increase the time it takes a user to get to the page, so I wouldn't be surprised with higher than usual drop off rate.

You should also compare mobile with desktop traffic and see how drop off rate changes. My feeling is that it might be mobile traffic that causes higher drop off rate.


over 5 years ago


Nick Fawbert

Sorry, I didn't see the rest of the comments

over 5 years ago

Brad Hess

Brad Hess, Copywriter at Sprinklr

Lots of very valid points made in the above comments...but what about bots? Much has been written lately in the ways of Twitter Bots.

Of your following, how many are bots that might be retweeting? Or similarly, users who comb their RSS feeds and RT?

At Sprinklr, we use a variety of measures to determine real social analytics beyond GA. But my main point of contention is this: while each post to my modest personal Twitter following might generate -- let's say, 15 clicks -- only 4 or 5 of those clicks are verifiable as human (using our language engine). This number is usually EVEN LOWER than the clicks number GA provides.

That said, it seems many clicks are falling through the cracks for you. Gotta love on vanity URLs!

over 5 years ago

Damion Brown

Damion Brown, Trainer & Consultant at Data Runs Deep

This really shows just how lacking utm tagging is as a reliable means of tracking complex campaigns.

While utms are better than nothing, it's still very very old Urchin technology that wasn't made for the way content gets shared today, across multiple channels and multiple devices. If everything was desktop PC across three or four completely isolated channels, I don't think this problem would present itself.

Have a look at Squeeze ( which is an advanced content-sharing / URL-shortening platform. Much more up to the task than GA.

over 5 years ago


Nathan Linnell

A likely explanation is that the click counts from the shorteners include bot and crawler clicks which will not show up in GA as visits. While the shorteners try to filter these out, there are likely quite a few that end up in the click counts.

On a separate note, here's a quick tip on how to easily filter for your Twitter data in GA. On the All Traffic report, use Source as the Primary Dimension and then add the following regex as an advanced filter (make sure to switch Containing to Matching RegExp)


Even better, create a custom report with the exact view you'd like.


over 5 years ago


Dan Bowyer

A generalist comment.

Suggesting that ALL analytics engines are a guide not an absolute.

I sometimes find that revisiting the same data set on different dates changes outcomes. So. With the best will and the best tools in the world, surely the figures can only be a guide?

over 5 years ago


Tim Aldiss

Fascinating. This is why I find it so important to use 3rd party tools like Tweetreach as a better metric for a broadcast channel like Twitter is reach.

over 5 years ago


Kieran Daly, Founder at Stone Circle Digital

" blog bounce rate is always fairly high, "

This for me in this post was really the most interesting point. I think the post is fascinating in the manner of the number crunching and data analysis (I love that area of Internet marketing). However when it comes to the crunch you are 'using' Twitter to both publicise and promote so...

Calling a spade a spade you hope that people will sign up for some of your paid offerings. I would dig deeper into your bounce rate over the raw stats of who/where/when. I am not saying disregard these as it is all part of the joined up thinking but if you see high bounce you may be wasting the energy in the wrong areas.

On the post I think that Google Analytics can provide a mixed bag of stats from time to time and there may often be some value in exploring he more powerful paid statistics packages that are out there. Especially for a high traffic blog like this. Have you ever done this and compared the effectiveness of the results?

Also there is an annoying habit in recent years of Twitter with people just retweeting links such as these to 'show authority' (I know snort) and people don't even bother their bums clicking on the link.

I plan to tweet this link as I read it liked the content and in fact have bookmarked it as some of the analysis techniques are interesting.

Have you conducted similar tests on Facebook/ Google+ and Linkedin? It would be interesting (for me for instance to just distribute a great post like this exclusively to one of these channels and look at the bounce rate from those locations to see how discerning the readers were there.

Thanks for the post enjoyed it while drinking the second coffee of the day!!


over 5 years ago


Rachel Buck

Great blog post, thanks. I agree with Neil, although there's bound to be some discrepancies in the numbers I'd also assume that many people simply re-tweet without visiting. You're all clever people and re-tweeting you makes us seem interesting and clever like you, right?

I always include anecdotal evidence such as conversations, comments, in my reports to back up the numbers, although this may be more difficult for you considering the number of followers you have.

over 5 years ago

Anna Lewis

Anna Lewis, Google Analytics Analyst at Koozai

Hi Matt, have you looked in the multi channel funnels report or the social report to see if the grouping that GA provides for Twitter is more accurate than the data manually grouped together? I know they are different kinds of reports and might not fit your attribution modelling but it might be that the data in there could make slightly more sense. But then again, I also find that direct contains social traffic which is frustrating across all reports!

over 5 years ago


Chris Roche, Digital Marketing Manager at New Gen Marketing Ltd t/a Instinctive Digital

I would say another thing to be careful about when measuring on analytics is making sure you are comparing the same figures.

In your example it shows the PAGEVIEWS and you are comparing this to the number VISITS but these are two different statistics. It looks like the actual number of visits would be more like 839 as indicated by the number of entrances to them pages.

over 5 years ago


, NA

The factor not considered in all this discussion is how reliable the Google Analytics tracking code is on mobiles. It requires javascript support, which is not uniform over all mobile browsers, and calls a number of external files, which may not get consistently obtained over mobile connections. I am becoming convinced GA is not reliable over mobiles, especially with regard to referrer attribution, but I lack the capability to run definitive tests. I have seen a number of people complaining the GA numbers for their mobile audience don't look right.

over 5 years ago


Stuart G. Hall

Or as my German ex-colleague at said after I reported another set of online community stats, "but what does it mean"? IOW "performance before reporting", as my ex-line manager said to me.

Whatever the stats do or dont say, how do they help 'move the needle' to meet business objectives, in the mix of all the other business analytics from the ROI of running a enterprise-wide multivariate analysis of your marketing data, to tracking Twitter referral visits to your blog? From a personal pov I like the fact that WordPress comes with such stats in the dashboard:-)

over 5 years ago

Matt Owen

Matt Owen, Marketing Consultant at Atomise Marketing

Hi all - first of all thanks for all the amazing feedback on this, certainly some things to think about, all much appreciated.

I agree that no analytics platform will provide perfect numbers, and if it's a few percent either way I can live with that, but this is such a large change that I think it's worth looking at more closely.

Anna - yes I do, and they seem to match each other fairly closely. My concern is that if granular measurement is wildly off then it undermines my confidence in the grouped figures as well - more to look at! :)

Thanks again all,


over 5 years ago

Andrew Hood

Andrew Hood, Managing Director at Econsultancy Guest Access TRAININGSmall Business Multi-user

One of the big challenges with this kind of comparison/reconcilation is that there's often more than one thing going on behind the scenes.

Another key consideration in this example is that if a user follows multiple links from Twitter within a relatively short period of time (specifically without taking a 30+ minute break from browsing the Econsultancy site) AND the traffic source remains consistent then that will all fall into one visit.

e.g. if I'm catching up with your twitter feed and click through on a bunch of links at lunchtime and they are all tagged up consistently (same source/medium or same referrer) then multiple clicks = one visit.

So a potential fundamental fruit salad issue (apples vs pears) on visits versus clicks.

The hypothesis of direct "stealing" from twitter is quite likely if end users are coming through from twitter apps (no referrer) and the utm codes are not in the URLs for whatever reason.

Incidentally, most of the twitter links from the Econsultancy twitter handle I've come across have utm_medium=feed/utm_source=blog in the final destination URL once unshorterned, so wouldn't be caught in any of the "twitter" nets described in the thread above...

over 5 years ago


Graeme cole

Social media measurement needs this sort of debate to progress, great comments.

Agree on the use of a single URL shortner and triangulating data.

However, with sample sizes and auto-URL extenders, the data will never match:

From "The numbers displayed are total decodes (not total click-throughs), which JavaScript measures on the page. Decodes can be caused by bots or applications, like browser plug-ins, which expand the underlying URL without causing a click-through."

over 5 years ago



Thanks Matt,

It is nice to feel less alone in the Analytics jungle.

The biggest difference I have noticed is between source 'twitter' and '' for other metrics like Avg. Visit Duration and % New Visits. I can't think what the cause is.

over 5 years ago



It might seem obvious (and apologies if this has already been covered elsewhere) but it's also worth bearing in might that probably only a relatively small percentage of followers will see any given tweet - especially if they're following lots of other accounts.

People often seem to forget this, but just because you've got 100,000+ followers doesn't mean any given tweet is going to be seen by 100,000 real people. If only 5% see the tweet and 5% then respond by clicking through to the target URL then a tweet from an account with 100,000+ followers might only generate 250 website visits.

over 5 years ago


Sebastian Crump, Freelance at SCRUMPH

Fascinating article and comment thread! I don't have anything technical to comment on, but just a usere behaviour point. I used to click on links in twitter to your blog posts, but I found they were 'too fresh' - so I now wait until the daily pulse to select the articles I find interesting (like this one) when the comment thread has had a chance to develop.

over 5 years ago


Bill Bean

This post is a great reminder that blogging isn't all about content marketing. The conversation in the comments has been particularly valuable. Matt, thanks for sparking a useful discussion.

over 5 years ago

Graham Charlton

Graham Charlton, Editor in Chief at SaleCycle

@Kieran the blog bounce rate averages about 76%, which is higher than the site average. However, some posts will have a rate more like 50%, and some higher, so we have some idea of what works and what doesn't in this respect.

I discussed this in more detail here:

Ideally, we want new visitors to come in through the blog, then hang about and explore further, though obviously this isn't always the case, and much depends on the topic of the post, and whether we give people enough further reading options for them to check out.

Btw - this article on Boagworld has some useful tips on measuring engagement on content sites which I'm currently looking into:

over 5 years ago


Joanna Chmielewska

Thomas Baekdal has a great post on the uncontrollable nature of a share and tracking challenges associated with it:

over 5 years ago

James Gurd

James Gurd, Owner at Digital JugglerSmall Business Multi-user

Hi Matt,

Great to see posts like this that provoke discussion. Understanding the data in analytics tools and knowing how to interpret it is essential and often not discussed enough.

My take on this is that using campaign tracking parameters is absolutely essential and to a granular level, not just medium, source & campaign. Use content/term to specify blog topic as well.

A lot of links shared on Twitter by Econsultancy don't use term/content.

This post could use:
medium - socialnetwork
source - twitter
campaign - blogtweets
content - socialmedia
term - [date published/tweeted]

I like to use campaign reports to compare against standard reports inc landing pages, accepting as Dan pointed out that the number are unlikely to ever match. But what's useful to monitor are the trends - if one report is showing a downward trend for a shared link and another an upward trend, then you can ask more questions.


over 5 years ago



Hi Matt,

Nice discussion and a methodical approach. I agree with points made by others that in particular that not all re-tweets are actually visits.

One minor point and I am not sure if you have used any other analytics software on your website too but Google Analytics has limitations on what it can track - basic things such as Java Script enabling or certain browser configurations might reduce your ability to track visitors using GA?

If you look at stats of people who have Java Script disabled there are about 2% in the US and about 1.5% in the UK - so this adds another level of approximation to any figures coming from Google Analytics?

What do you think?

over 5 years ago

Antoine Gay

Antoine Gay, Sales Manager at Tag Commander

Interesting article, important to note that :
- Sampling in GA is up to 500K visits
- Other analytics solution don't sample but in that case we are talking about smaller numbers
- Do you have another tools to compare on a similar perimeter? (I see speedtrap and quantcast)
- Multichannel and attribution are worth looking into (focusing on all but last click)
- comparing clicks and visits make no sense, you might want to track your links with a redirect to compare clicks in twitter and clicks in GA
(sorry if I've repeated what might have been mentioned)

over 5 years ago

Andrew Smith

Andrew Smith, Director at eschermanSmall Business

Great post. An excellent example of many eyes on the problem, etc.

So what constitutes a good/bad/indifferent click through rate for content shared via a Twitter account such as @econsultancy? With 100,000 real followers, a natural instinct is to assume that traffic from this source should be of a certain level.

But as others have already pointed out, there could be a number of factors that would reduce the real reach of your Tweets (and thus impact on click throughs from Twitter)

Of your 100,000 or so “real” followers, how many are actually exposed to a Tweet at a given time?

Bruce Daisley of Twitter UK said last week at the Social Brands conference that 40pc of Twitter users never Tweet - if they are only in viewing mode, perhaps many won’t even bother to click what they see.

Followerwonk ( seems to suggest that the peak time for your followers activity is around 2pm UK time - with 6.2pc - so a Tweet you send out at 2pm would (according to Followerwonk) be exposed to only 6200 of your real followers. Depending on how many people they are following, your Tweet might get buried in their general home feed. (As an aside, using Followerwonk’s new social authority metric, the tool claims that nearly 69pc of @EConsultancy’s followers have a social authority score of less than 10 - would have that have an impact on whether or not these followers would actually do something like click a link?).

I know various people have had a stab at the shelf life of a Tweet which ranges from 18 mins to 2 hours. Even so, that would at best expand the visibility spectrum to around 12000 at peak times (assuming Followerwonk’s figures).

So lets say you get 600 clicks for a Tweet. If your real reach at 2 - 3pm is 6000, then you have CTR of 10pc - not too shabby.

I’ve looked at this before - take the New York Times - ostensibly, it has 7.5 million followers (ignoring the fact that a big proportion of those may be fake or inactive), but the click throughs on its Tweeted links sometimes barely make it into three figures. Here are the stats for a typical story - 815 clicks from 7.5 million followers. On that basis, I’d say @EConsultancy CTRs (given your follower numbers) are pretty stellar ;)

Here are a few posts I've made on this subject in the past:

Before anyones says it, of course, we need to treat any of the above with caution - how accurate is Followerwonk, etc.

But as others have pointed out above, other impacting factors related to GA include visits from mobile (and a big thank you to Joanna - that Thomas Baekdal post is a corker) - if those mobile visits are being dumped in Direct (and if 80c of UK Twitter users access via mobile), then you could well be losing a lot Twitter traffic numbers due to this.

So what to make of all of this? Well, the level of commenting here shows that this is a big issue - and we are all people who take a big interest in it and will know more about the ins and outs of Twitter/GA etc than most. So if we are scratching our heads then think of everyone else.

Having said that, clearly we are all focussed on the problem - and inspite of our concerns over GA accuracy here, surely the insight we are able to gain today is WAY better than what could have been achieved even a couple of years ago. That must be something to be cheerful about rather than pessimistic.

And finally, should we be concerned that the numbers of clicks as a proportion of the theoretical reach is low? Surely we all agree that quality of traffic is more important that quantity. Using attribution analysis in GA time and again shows the direct and indirect contribution of social media - the numbers may be lower, but how often do you find that social media visits are a higher quality (stay longer, higher propensity to purchases, etc) than say natural search visits - even if the latter sends lots of raw traffic?

Thanks again for bringing the issue out into the open and your honest appraisal of it - the more of us who work on the problem, the quicker we will have a solution that will beneift us, our clients and the organisations we may work for.

over 5 years ago

Matt Owen

Matt Owen, Marketing Consultant at Atomise Marketing

Thanks Andrew, excellent point - big fan of followerwonk and also socialbro's audience analytics.

According to my 'averaged' figures (from a few different tools, Econsultancy's largest potential Twitter audience is online at 2-4pm, Weds and Thurs. This makes sense as we'd have both our UK and US audiences online and these are high social activity times. Up to 26% of our audience is on Twitter at those times, but as mentioned, that doesn't mean that they'd be exposed to a tweet, and of course, we don't always tweet at those times.

This is something of an aside obviously, although we can and do schedule content to reach audiences at these times, the figures above relate to actual clicks vs measured traffic, so while a larger audience might effect the number of clicks (and retweets), and analytics would pick up a correspondingly larger number of visits, it probably wouldn't affect the accuracy of that measurement. We'd know we got more, but we still wouldn't know the percentage of untracked visits.

For the record, Twitter tells us that around 2-400 people usually click on each of our tweets (that contain links to the blog), and we tweet around 25-30 times in a 24 hour period, so we should expect at least 6-7,000 blog visits from twitter per day.

In reality it's much higher, that figure is purely people clicking links we have sent from our account. Lots of people retweet and share from the article page itself, from their own accounts, from each other, and lots of them DO click, we know that from the traffic we CAN track, but how many more can't we measure... back down the rabbit hole we go... ;)

Currently running a lot of tests so expect a follow-up soon, and thanks again for all your comments!

over 5 years ago


Benjamin Mangold

Great to see such a great discussion about social media attribution!

I totally agree with Neil, in that the assumption here is that people click on links within Tweets. This is a pretty big assumption and I would start by focusing on CTR (Click Through Rate) of your posts. It is also worth comparing Twitter to your other social channels to see if there is any correlation in click. For example, you can head over to Facebook Insights to compare clicks and CTR of your posts. This might not be perfect, but you can begin to compare performance.

As Dan pointed out, you aren't consistently using UTM tags on your social posts, so I would suggest you setup a process to ensure consistency. Looking at a few of your links there is also incorrect tagging, in that you have used utm_medium=feeds&utm_source=blog for one of your links, so you are causing yourself accuracy issues there. Unless you are going to be 100% on top of tagging links I always recommend leaving organic (or free) posts without campaign tags. This ensures that they will show up in your default Sources and Social reports. Then you want to ensure you use utm_medium=social&utm_source={domain of social network}&utm_campaign={campaign description}&utm_content{ad description} for more on this you can checkout this video on YouTube -

When you are adding links to organic posts you can use a shortener that allows you to see where the link is being clicks, for example, the Google URL Shortener ( allows you to see the different referrers that are resulting in clicks. This can then be compared to your Google Analytics data. But you really want to be using both the Social reports and the Sources reports (under Traffic Sources) for analysis. The social reports aggregate data for individual social networks which can be a simpler way to see data than the Sources report.

The other thing to remember is that digital analytics is not perfect, get over it and use the actionable data that you have available! Define metrics that enable you to take action and see value where it is available, this includes metrics for branding, engagement and conversions. Measurement should be tied to action and optimization, so by comparing social networks and channels with actionable metrics you will be able to improve performance over time.

over 5 years ago


Steffen Kastner

It is also worth noting that with the release of iOS 6 the majority of your traffic coming from iOS devices will only show up as direct traffic:

Since no Google search referrers are forwarded any more my bet is that your biggest share in direct traffic will be from Safari, followed by Android Browsers. Android users are usually logged into their Google account and will also provide no referral information to Analytics.

It might be possible that referral information ist not only being stripped from search traffic but from all secured sources users like a twitter page where a user is logged-in.

It might be a good idea to have a look at the figures before the release of iOS 6 (Sep 19th). In my Google Analytics account I can tell the date by looking the graphs. :-)

over 5 years ago


Oddny Edwards, Marketing Manager at CuCo Creative

I would just like to thank you for a great article and all the comments. I have been ripping my hair out trying to figure out why my numbers weren't tallying up at all.

Why is it so hard to have a simple analytical source that indicates where the clicks come from.

Does anyone think that this is a way of getting all these analytical platforms to convert us to being paying members? At that stage they will promise to show you directly where the clicks are coming from and their conversion? Or have I been doing this for too long and am simply far too sceptical?

Thanks again!

about 5 years ago


Robert Rogge, CEO at Zingword

Hi Oddny,

I read through a bunch of comments, and I'm forever skeptical of Google. But I think that the reason isn't to make you pay for premium Analytics, it's to make you buy more Google ads.

Already they are removing most of the keywords from your organic search results reports, which not only obfuscates things for SEOs, it also makes businesses turn to Adwords because at least it can be accurately measured.

By misreporting social media traffic, Google would stand to gain in revenues from Adwords. Even in this post, he talks about how it's hard to request your budget. If your job is on the line, are you going to bet on social media that you are *pretty sure* about, or are you going to bet on Adwords that you are *totally sure* about?

I'm not necessarily accusing Google. But they stand to gain from this, and based on some of the suits they've lost here in Europe, it would not surprise me.

about 4 years ago

Save or Cancel

Enjoying this article?

Get more just like this, delivered to your inbox.

Keep up to date with the latest analysis, inspiration and learning from the Econsultancy blog with our free Digital Pulse newsletter. You will receive a hand-picked digest of the latest and greatest articles, as well as snippets of new market data, best practice guides and trends research.