Category: social networks

28 March 2013

More projecting, less sharing

In 2008, the sociogeek study asked respondents whether they would post specific pictures on their profiles. The results – partly shown below – are far from surprising: the more compromising or private a picture is, the less likely it is to be shared. Despite many claims of the contrary, Sociogeek showed that social media users are deeply in control of their image. They do not show who they are, but what they want others to believe they are.

Nobody shares, everybody projects.

^{Results of the sociogeek.com study. The % show what proportion of users would feel comfortable sharing the corresponding image.}

Now another factor will likely accentuate this projection phenomena: the fact that social media data is increasingly used to quantify and measure who we are. It is already happening in finance, where Facebook data is used to calculate whether a person is creditworthy:

Facebook data already inform lending decisions at Kreditech, a Hamburg-based start-up that makes small online loans in Germany, Poland and Spain. Applicants are asked to provide access for a limited time to their account on Facebook or another social network. Much is revealed by your friends, says Alexander Graubner-Müller, one of the firm’s founders. An applicant whose friends appear to have well-paid jobs and live in nice neighbourhoods is more likely to secure a loan. An applicant with a friend who has defaulted on a Kreditech loan is more likely to be rejected.

Link on economist.com

There is an interesting cat and mouse game looming on the horizon:

Social networks need users to be as truthful as possible, as they live off advertising. To target ads effectively, Facebook needs to know what people really like, who their friends really are. Social networks want us to share, not project, as the relevance of their ads depends on the quality of the information they have.
When Facebook data becomes a way to calculate credit ratings, users are pushed to manipulate their data, befriending people from rich neighbourhood, liking luxury brands, anything they think will make their ratings go up. Users will be pushed to project an image, and not share the reality of their life.

Overall, we see two very contradictory forces emerging, one pushing for more transparency, the other for more opacity. What will happen once users start to “strategize” their facebook presence? How will data scientists separate a true like from one made with a specific goal in mind?

Here is the big question if you are Facebook: what do you gain from giving access to users’ data, and what do you lose? What if allowing the calculation of credit ratings was offsetting the possibility to target ads effectively?

That is a key question I would really carefully look at if I was Mark Zuckerberg, to make sure Facebook doesn’t end up being a network of shadows geared at getting Klout perks and other benefits.

2 August 2012

Column in the September 2012 edition of Wired UK

Here is my column published in the September 2012 issue of Wired UK, the best magazine I know to keep an eye on what is happening in the fast-moving world of innovation and new technologies. Subscribe if you haven’t already, 24 issues for only 48£ (am doing the sale because I really believe if you are interested in the future you should be reading Wired UK, it’s even better than the US version).

How not to be a product on Facebook

If you have shares in Facebook, you probably believe that this platform will sooner or later command a dominant share of the surging online-advertising market, estimated by MagnaGlobal to be worth $103 billion (£65 billion) by 2015. It’s a reasonable assumption, really. Advertising on the social web is an enticing prospect. A place where people tell you who they are, what they like, what they listen to, what they watch: in short, everything about everything. In 1960, if you had told Madison Avenue professionals that their spiritual heirs would have access to endless stores of detailed personal information on billions of people, they would have licked their chops. Or perhaps they would simply have wondered if the end of democracy was imminent.

Indeed, social networks such as Facebook are treading a fine line, balancing users’ privacy on one hand and the need to mine data on the other. To deliver ads that are relevant, advertisers need to know as much as possible about their audience. But they also need to show that they are using information in a reasonable and responsible way in order not to scare people off. An ad triggered by a recent iTunes download is fair game. Suggesting a diet because I sent an email with the words “chocolate” and “fat” is harder to accept.

So far, social networks have operated in take-it-or-leave-it mode. You either accept their terms of service wholesale, or you walk away. There is nothing in between. Enter Privly, “an open-source community project for taking back control of personal information”.

Read full column

23 June 2011

The invisible rule of proportionate attention in online communication (and why social technologies are not magical)

I send a lot of emails. I post a lot of messages on my blog or on Facebook.

One thing I have noticed over the years: there is an invisible rule that seems to reign on the online world, regardless of the medium: the more care you put in a message, the more chance there is it generates an answer (email) or interaction (social networks).

Take email. When you send a newsletter, if you get 50% of people opening your message (as we do at Lift) you can be pretty satisfied. The industry standard is more around 20%. That is what you get for sending messages that have not been specifically written for the recipient. They feel that, and have no pressure to answer whatsoever as it has been sent to thousands of people.

In a typical one to one communication, answer rate is probably closer to 95% as long as you write to people you know, and who are at the same “level” than you.

Now for my editorial job at Lift, I get to invite pretty busy people as we try to convince them to join us for the conference. We don’t always succeed in having them, but at least I get around 80% of answers to my messages, positive or negative. I get this answering rate by carefully crafting my messages to make the recipient feel I value him or her, as I invest a lot of my time in reaching out. If I send a quick message, it is likely I will not get an answer. If I take time to research the person I am contacting, find out what their recent projects are, add a few personal sentences about the city they live in, the chances for a response get much higher.

My point here is that it seems that electronic communication is not totally deprived of context. When you talk to someone, your body language gives hints of how you feel, and influences the answers you get. In electronic form, these implicit messages can also be conveyed. I care about the discussion we’re having, I’m willing to invest time in reaching out to you. That matters.

I noticed the same happens on my blog and on Facebook. On the blog, articles where I simply pass a link (as I often do to set them aside for my personal archive) receive little feedback, while longer and more personal articles generate more comments. On Facebook, I have an even more tangible proof. For a long time, the Lift page was managed manually. I would replicate each article carefully, adding a custom message different from the title of the news I was pushing to the community. As soon as we installed an automatic app (RSS graffiti) to republish articles automatically, the number of interactions almost halved. It was the same content, but our followers felt we were not putting as much energy in the process of pushing the information to them. They felt less engaged, maybe less cared for, and the number of interactions dropped.

That’s why social technologies will never be magical. They promise us more personalized interactions with followers, as we know who they are. Truth is, mass updates will always have a different feeling from a message written specifically for a recipient. Nobody can escape the time consuming task of writing personal messages. And if you have 10’000 fans, that will take a while.

29 March 2011

Online, quality beats quantity

Online used to be a world of “the more the better”. Search engines would brag about indexing X zillion of pages, the race for LinkedIn contacts was raging between users, and even professional sites like Alibaba would base their communication on the fact they allowed to reach thousands of suppliers in one click.

For early adopters, every single piece of information published on the network was a small victory. Each page indexed by Altavista was one more step towards the society of information that we were trying to build.

This was the old world of megabytes. Not the world of 2011, measured in exabytes. Tons of information have now created noise when we increasingly need relevance. From a world whose problem was to add information, we now enter a world where the problem is to find which one can be ignored, hidden, or deleted. Let’s take the three examples again:

Do you really care if Google indexes one or two more billion pages? No, you care about the top 10 results. The challenge is not to index 90 million pages containing the word “bank”, it is to hide the 89.99M that are not relevant to the current context.
Users are coming back from the “more friends is more fun” mantra. I see people remove or hide friends, some now cap the number of contacts to a “few” hundred. But the key is advertising: once social advertising happens (whatever it’s form), more friends will likely mean more ads. “De-Friendization” will then accelerate. And what is the point anyway, when we all lost the followers race to Lady Gaga anyway ;)
If you have to find a supplier, would you rather have a lot of offers, or the right ones? Sites like needeo work with selected suppliers, not “all the world’s suppliers”, and in that case quality and trust will always beat quantity.

Another sign this trend is here to stay can be found in services like Path which, in their DNA, embed the fact that you can not keep in touch with more than 50 people. Also Beluga (acquired by Facebook), Brizzly, the trend is now to launch closed group apps, to capitalize on the fact that it will be easier to monetize systems based on quality relationships than on a lof of relationships. The data mining will be easier (less data to make sense of), the social ads will be more effective (users are more likely to click on a recommendation coming from a close friend than from an acquaintance), and it will be possible to create real trust between the users and the system, with no fear of privacy boundaries being crossed.

From a world of quantity, we now live in a world of quality. The key is not to have a lot of signals, but to have the right ones. Social networks make it possible as long as they don’t encourage us to have lots of friends, just the right ones. Do you now better understand the 75b$ valuations of Facebook?