Yesterday, SpotOn Public Relations released a widely-cited report that detailed the amount of Twitter users in the Middle East and North Africa. We published two articles based on the figures; one by media reporter Keach Hagey, "Middle East Twitter users sees rapid growth", and by technology reporter Tom Gara, "Facts show Twitter yet to lure Middle East users".
In short, the numbers gave a brief introspective view at how the micro-blogging service has grown in the area, while ensuring that if there's any local companies out there looking to do some "internet marketing", it looks like SpotOn knows their audience.
But I was skeptical. I spent over five years of my undergraduate education analysing some pretty intense statistics and I just didn't buy SpotOn's methodology, only cited in their press release as: " based on the agency's survey, tracking and analysis of registered Twitter accounts across the Middle East and Africa region."
To confirm these numbers, I turned to Sysomos, a Canadian-based start-up, who hold a patent-pending technology born from a university research study on analysing the social web in real-time. The company made its mark among the Twitterati after publishing a fairly comprehensive global study on the usage of Twitter, full of fancy graphics and all that.
It apparently took the brains over at Sysomos quite a bit of time to process the data. I received the figures from Sysomos late last night, past our paper's deadline, from company spokesman and Twitter expert in his own right, Mark Evans.
The results comparing both firms' figures are after the jump:
What can we see from above is that while the numbers are close, they aren't perfectly spot on (pardon the pun). Although some areas such as Tunisia represent a 1 per cent difference, there is a 54 per cent discrepancy with Saudi Arabia's numbers. Closer to home, Sysomos found more than 700 more Twitter users in the UAE than SpotOn's numbers. Even the final total has a 10 per cent difference of about 1,400 users. Given the overall population of the data, that's quite the tolerance, if you're a stats geek.
What explains the discrepancy? I'll let SpotOn comment below on how they tracked their data. Meanwhile, Sysomos taps into Twitter's API (the data "firehose" itself) to yield is results and are funded by the venture capital arm of the Ontario provincial government. I think they're legit.
In any case, the numbers reinforce my blog colleague's view that Twitter just isn't that important here in the Middle East. As he puts it:
But what it has not achieved, and seems unlikely to do, is change
the dynamics of a Middle-Eastern internet culture that has shown a
clear disinterest in the kind of open, public self-expression that
Twitter is all about.
I think the point that Spot On was trying to make still stands, irrespective of what the numbers are exactly. The variation is so little in the bigger picture that I don't feel the new results show anything groundbreaking, or deserve to become a new story.
It's very interesting to see the differences between Sysomos figures and ours. To be clear from the outset, Spot On started tracking Twitter usage in the Middle East to get more insight into adoption trends in the absence of any published statistics for the region. However, we're big fans of Sysomos and recommend them often these days for global Twitter statistics.
I think what we need to bear in mind here is that neither the statistics from Sysomos, nor Spot On, are precise. There is currently no 100% accurate way to measure Twitter adoption in the region. We're not claiming our statistics are precise and I wouldn't expect Sysomos to do so either. Our figures are our best estimates based on research of registered user profiles, annecdotal insights and our experience in the region. The two pages of figures and charts which you used as the source for our figures states "All statistics are derived from the research and market estimates of Spot On Public Relations."
I'd like to be able to say in all honesty that Spot On's figures are all more accurate than Sysmos, but the likelyhood is that some of our figures are closer to the truth and some of the figures from Sysomos are closer. Overall, we're quite comfortable with our figures, but would love to improve on them.
Spot On PR has been a strong supporter of, and contributor to, social media communities in the Middle East because we believe in the value of the community model. Your first reaction to the figures we presented was that they were ‘fishy’ (via Twitter)– a reaction that is borne out by the tone of your blog post. Yet at no time did you bother to actually talk to anyone from Spot On and even attempt to explore our rationale, ability to talk to these figures or sincerity in offering this data to the market.
Surveying the Middle East Twittersphere is not as simple as capturing data and number crunching the results. There are plenty of anomalies that can skew figures, particularly when the region is in a very early stage of adoption and the number of registered users is very small. In these markets a factor such as the trend of users changing their location to Iran can be significant. As can irrelevant data from larger, more developed markets.
To pick a country example, the Sysomos figures indicate that Lebanon has 1,255 Twitter users. This is 621 users more than Spot On's estimate for Lebanon. Not a huge margin in itself in terms of numbers of users, but as a statistic it's a 49% difference. Spot On has been tracking Twitter usage in Lebanon since last year and we've seen the numbers of users grow from 60-odd in December, to over 600 in July. Reviewing Twitter user profiles that claim to be from Lebanon reveals that just under 50% of those users are actually from the USA, because there are towns called Lebanon all over America. Lebanon, Pennsylvania alone has more than 100 Twitter users who have 'Lebanon' in their profile (although many do not mention it just to avoid confusion with Lebanon in the Middle East). Kentucky, New Hampshire, Ohio, Missouri and Tennessee account for another 200+ US Twitter users that are from 'Lebanon'.
Similar anomalies skew figures for many of the countries in the table and whilst there are towns in the US with larger Twitter communities than whole countries the Middle East, regional stats are easily skewed. We have the utmost respect for Sysomos, but it just might be that, in this case, Spot On is closer to the market than they are and has more time to spend on tracking these smaller markets. We'd love to see Sysomos tracking, analysing and publishing data from the Middle East more frequently and hope that they do so. In the meantime, we've so far found our data quite useful for tracking regional trends.
@Carrington Completely agree with you that measuring the number of users on Twitter precisely is impossible. The numbers that Sysomos reported are simply based on what users disclosed based on our collection of data (and obviously, some lied, some did not disclose, no one has complete Twitter data, and so on).
We however did take in account all the possible steps to ensure that the numbers are as accurate as possible. For example, we specifically considered only the Location field eliminating mentions of Oman in bio/other fields. We also did use our geo-parsing technique which is reasonably accurate in separating locations with same name in different countries such as lebanon. Both English and Arabic language location fields were parsed. Obviously, there always are a number of factors that can make numbers in such studies fuzzy (sampling techniques, estimation methodology etc).
I think the exact numbers are less important, and the overall message is significant: Twitter is still in its infancy in middle east reason.
Its quite interesting to note that the estimate given by Google Adplanner (which takes the info from multiple sources including comscore)puts the total users in UAE at 120K and the Saudi base at 100K....
Would be keen to hear everybody's thoughts on this ....
I can't comment in detail on Comscore's figures, you'd need to ask Comscore for that. However, my understanding is that those figures reflect web traffic (I'd expect unique visits) not numbers of registered users. However, it does provide another reference for the impact that Twitter is having on the Internet and its relevance to SEO. These days Twitter user profiles appear very high up Google search results when searching for people and / or companies and users tweets are appearing more and more in key word searches. Eg. search for - MENA Twitter - now and you'll see tweets being returned first in the search results.
i am trying to find twitter stats for israel- do you have any information