Category Archives: Social Media

The Realities of Twitter Democracy!

In September 2011, as the then editor of Dataquest, I wrote an editorial, The Opportunities and Threats of Facebook Democracy. While Dataquest was one of the first publications to do a cover story on how social media was effectively used in the fight against corruption earlier that year (in April), and had celebrated the new power that social media had given to the common people, in this editorial, I had warned against attaching too much importance to the voice emanating from social media by the leaders and policy makers. My reason, of course, was the low penetration of social media. In a large diverse democracy, jumping into conclusion based on what a small section of people belonging to a particular socio-economic section say, was a potentially dangerous and suicidal thing to do, I argued.

The reason I called it Facebook Democracy was that a lot of the campaign by India Against Corruption was actually carried out on Facebook. It was the main mobilization platform.

Since then, Twitter has been used by politicians very effectively to drive their messages. Many politicians and political parties have taken professional help for that purpose. All of us know the power of #pappu and #feku campaigns. While penetration of Twitter is still miniscule compared to the size of  Indian electorate, some politicians have managed to have a very large fan following, going up to more than a million. And there are at least ten Indian politicians on Twitter who have more than a lakh followers. Considering that not more than 100 million Indians are online, those numbers are not unimpressive.

Unimpressive they may not be. But as it turns out, most of these followers are fake.

Social media management platform maker, Status People, actually provides a way to check your (and others’) fake followers. I actually checked out the the fake followers of the top ten Indian politicians on Twitter by number of followers and checked how many fake followers they have. 

And can you imagine what the average looks like?

It is 59%.

That is, as much as 59% of the followers of these politicians on Twitter are fake. And typically, the bigger the number of followers, the bigger is the percentage of fake followers, though there are small exceptions.

Here is the table.

Politician Twitter Handle Total Followers Fake Followers (%)
Shashi Tharoor @shashitharoor 1756468 62
Narendtra Modi @narendramodi 1560092 65
Dr Manmohan Singh @pmoindia 538323 55
Sushma Swaraj @sushmaswarajbjp 447766 52
Arvind Kejriwal @arvindkejriwal 314614 54
Omar Abdullah @abdullah_omar 274937 54
Subramanian Swamy @swamy39 165408 42
Ajay Maken @ajaymaken 151118 55
Derek O Brien @quizderek 149448 38
Varun Gandhi @varungandhi80 118728 52

The numbers are as on 1st May 2013

And here are some realities.

  • Narendra Modi, the potential PM candidate of BJP, heads the list in terms of  percentage fake followers, with 65% of his followers being fake.
  • As many as 8 of the 10 in this list have more fake followers than they have genuine followers. Derek O’ Brien and Subramanian Swamy have the lowest percentage of fake followers in this list.

What Does This Mean?

This, of course, does not suggest that politicians are doing something deliberate to create fake profiles/followers. And since there is not much to choose between different parties, it is not a political statement that one is making. In fact, many politicians themselves will be shocked to know this.

For that matter, there is not too much of a difference between politicians and other celebrities when it comes to the percentage of fake followers. I did check that for a couple of them. In case of Amitabh Bachchan, 73% Twitter followers are fake. For Shah Rukh Khan, that number is 70%.  But in case of celebrities, it is a reaching out to the fans, so it does not matter how many fans follow them.

For politicians too, it is a great platform to get their message across, engage with media and at least a certain section of people, who are using this medium. The problem begins, when, their PR managers try to make us believe that they are great leaders because of the large fan following. That is when we get it completely wrong.

In fact, fake followers is just one part. The above platform, Status People, also measures how many of the followers are inactive. For each Twitter profile, it divides the followers into three parts: fake, inactive and good. When you take just those followers that it terms are good (who are real and active), the total followers number drops drastically.  Here is the above list of politicians with their “good” followers.

Politician Twitter Handle Good % Good Followers
Shashi Tharoor @shashitharoor 10 175647
Narendtra Modi @narendramodi 10 156009
Dr Manmohan Singh @pmoindia 16 86132
Sushma Swaraj @sushmaswarajbjp 16 71643
Arvind Kejriwal @arvindkejriwal 14 44046
Omar Abdullah @abdullah_omar 13 35742
Subramanian Swamy @swamy39 21 34736
Ajay Maken @ajaymaken 12 18134
Derek O Brein @quizderek 22 32879
Varun Gandhi @varungandhi80 13 15435

The numbers are as on 1st May 2013

So, in effect, Shashi Tharoor’s active are just 1.75 lakh Twitter users, not 1.75 million.  The prime ministerial candidate Narendra Modi has just 1.5 lakh followers, not 1.5 million. Varun Gandhi just has 15,000-odd  active followers.

In short, these numbers denote their actual sphere of influence. Except for Tharoor and Modi, these numbers are in thousands; in a country of a billion. And when you combine this to the fact that Twitter reaches only a certain class of people, it follows quite logically that extrapolating the influence/opinion of Twitter to the real world is not a great idea. Not yet.

Advertisement

1 Comment

Filed under Digital Economy, New Governance, Policy & Regulation, Social Media, Technology & Society

Whatever Its Algorithm, Klout Must Fix Its Basic Technical Issues

There has been a lot of debate about Klout, its score and its relevance. While some are addicted to it, many others dismiss it outrightly. Most of the criticism has been about the way it measures influence or its algorithm;  its non-transparent mechanism; and its scant respect for individuals (you have a Klout score the world can see and all your information with Klout, even if you have never heard of it).

There has been some good articles on what exactly is wrong with Klout. Here are a few. Why I quit Klout, Why You should too… and The Problems with Klout. You can find plenty of such posts and articles and you may tend to agree with many of those concerns. Others argue that it is still experimenting and should be given some time before it is dismissed. This is especially true about the criticism Klout draws about its presentation of the topics of influence, which sometimes are more than funny. I myself am supposed to be influential about  games. I still cannot figure out ABC of games that my six year old son plays so dexterously.

But, most of the criticism about privacy, transparency and efficacy of its algorithm are subjective. The disastrous measurement of topics of influence, which many argue, is a proof of non-efficacy of its algorithm, can probably improve as it is something that is a first in the world.

But what I cannot digest at all is that something that claims to measure the influence of the entire populace of the world is struggling to get some of the basic things in place.  I am talking of its interface with Facebook. While Twitter updates and interactions get updated in 48-72 hours (And you think that is too slow?) the Facebook interface is pathetic. And I am being polite. Sometimes, it takes a FB interaction 7 days to show up as moments in Klout, sometimes it takes 10 days, sometimes more. As of today (9th November), my last FB interaction that shows on my interaction page is of 25th October and that shows on my moments page (which presumably goes to make up the Klout score) is of 23rd October.

What is more, it is not a complete list. Anywhere between 20-50% of those interactions never show up. After I double-checked that they were public interactions, I wrote to them and they admitted that, it was a problem. “We are working on this issue currently and hope to release improvements soon,” I got the reply on 9th October. That is exactly a month back. I am not being judgmental on the time they are taking. But what I am absolutely worried about is that in the meantime, they continue with presenting the score to the world, which by their own admission, is not based on correct data. One can keep arguing about the algorithm. But there is nothing to argue if your data captured itself is not accurate.

In the same mail, they tried convincing me that it is only display of moments  that is an issue and the FB interactions are still being captured for calculating the Klout score. When I wrote back refuting this claim, I got a single line reply that they are investigating it and “have taken note of your account”. This was on 10th October and nothing has happened. In the meanwhile, I have tried disconnecting and reconnecting Facebook and still have faced the same issues.

The problems that potentially arise from this are multiple. One, the Klout score is based on only partial and haphazard data of users. That puts a question mark on the basic offering itself: the score.

The delay also is an issue. If there is a uniform delay in Twitter, FB, and other networks, one can still justify it saying it is a delayed feed. But imagine trying to create score from your activities and interactions on Twitter on 1st, on FB on 15th and Google Plus on 30th and combining them to create a score. What will that denote? And how will you relate that to any offline/online events? It becomes a useless number.

While many dismiss Klout, I am still of the opinion that it should be given a chance. But rather than trying newer things and fancy toppings, it must get its basics right. There is no excuse for basic technical issues. I would say proceeding further without getting its data integrity right will be a dangerous path for Klout.

 

Leave a comment

Filed under Digital Economy, Social Media