Category Archives: UI

How social networks should take advantage of sheep mentality in targeted advertising

Facebook now has more than 400 million active users. 50% of these log on to the site at least once a day. 3 billion photos are uploaded a month. These figures are staggering. Facebook now drives more traffic to news sites than Google, according to GigaOm. Farmville – a small Facebook-based game which depends upon social interaction – actually has three times as many active users as Twitter.

In fact, some people may be starting to think that Facebook is the Internet. A few weeks ago, a post on www.readwriteweb.com ranked very highly for the search term “facebook”. As the comments on that page demonstrate, users did not understand that this was not Facebook. They complained about the new format and the missing login box, whereas most of my readers (I imagine) would certainly have realised the difference, and would have typed facebook.com into the location bar of their browser initially in any case. Some of us may take a cheap laugh at this, but the sheer volume of comments there (not taking into account those that didn’t comment) demonstrates that this is not unusual. Many people search for what they want, and dangerously assume the first result is what they are looking for. A friend of mine’s wife never browses to youtube.com – she searches for youtube and clicks the first link.

Facebook likely knows more about you than Google does. There is widespread concern that Google could be relating our searches and browsing history together, and that there are privacy implications for this. But, for the huge amount of people that use Facebook as their portal to the Internet – not only do they know what we are sharing, clicking on, and interacting with – they also know our personal details that we give to them willingly, and even who our friends are and how often we interact with them. At a recent event I attended, an employee of Facebook told me they have enough data now to be able to fairly accurately predict when a couple who are interacting will change their relationship status to show that they have got together, or broken up.

Surely this amount of data gives Facebook the ability to start the really targeted marketing and affiliate network advertising that I have been hoping for? I do not think it is hard to conceive that at some point in the near future, advertising will no longer be an irritation. If the advertiser knows enough about me to present me with advertising that I genuinely am interested in, it becomes a service. For advertisers too, presenting an ad to a group that are more likely to be interested in their product enhances their brand rather than tarnishing it, saves costs on unnecessary blanket marketing, and ultimately increases ROI.

This has been discussed many times, and Facebook and Google (and others) are already trying to achieve it. Google’s effort is based on keywords within the page and is not yet as intelligent as we would like it to be. At best this can have funny results, and at worse it can be very offensive. Some examples are here.

Facebook makes a better effort. Since it knows the bands I enjoy, it tells me about tickets and upcoming concerts, with affiliate links. It also gives me the ability to effectively vote ads down that do not interest me, for whatever reason. But, Facebook have the potential to take advantage of sheep mentality (also known as group think, or group solidarity) – the mindless following of peers that the vast majority of us subscribe to. There have been many studies on this, with the outcome being always that we are more likely to subscribe to the same opinions, use the same services, and buy the same products that our peers and (to a lesser degree) the general populous do.

Facebook knows my interests. It knows who my friends are, and it knows of them who I hold in greater esteem based on the status updates and other shared items that I “like” or respond to. If they were to surface ads based on how many times they have been interacted with by my friends, or products that they know my friends and peers have purchased or are interested in, there would be a clearly stronger chance that I would also be interested in these things. If my friends that I hold in highest regard are treated as more important in these algorithms, the effect would be stronger still. The events application within Facebook is the best example of what I am getting at, but seems to miss out on the affiliate and advertising potential.

It is one thing to present me with artists that they know I am likely to like based on artists that I have listed as my favourites and other people’s listening preferences who also like the same artists; the value of my circle of friends strikes me as being more valuable than this. Some examples.

  • “10 of your friends including John, Joe and Mary all bought this item.”
  • “6 of your friends that listen to Nirvana also like Pearl Jam.”
  • “12 of your friends will be attending Reading Festival 2010, click here to buy tickets.”

Each of these examples would of course be accompanied by a link to buy the product or service in question, or useful services that relate (for example hotels and restaurants in Reading). This kind of advertising is not only useful, since it gives me information on people I am likely to care about, but it also offers a clear and strong revenue stream based on affiliate links and purchasing opportunities. I believe that using a user’s social network as data to base targeted advertising on is a logical next step.

What do you think?

Update: Andy Beeching pointed me to this article, which describes your circle of peers and the value of individuals within in as the “Social Graph”. It makes for interesting reading.

If you have to make your own dropdown list boxes…

The dropdown list box, or select box, is a massively useful UI tool. It presents a series of predefined options in a format that doesn’t have to eat up all our screen estate. Unfortunately though, we just don’t have the ability to style this that we would like. Or, sometimes none at all. Since this is an OS control, the OS takes control – and apart from the visual implications this is no bad thing. If you absolutely have to have it match the design for the rest of your site though, you have no choice but to implement it yourself. You can achieve this either via JavaScript or Flash. Flash is probably overkill if the rest of your site doesn’t use it, so if you use JavaScript I would heavily advocate the use of progressive enhancement to start with something that works (the select tag) and build something nicer on DOM ready.

If you’re recreating something the OS does though, please try to do it as well as the OS does!

Here’s a useful list of behaviours for a dropdown list box which might seem obvious, but I have seen overlooked so many times.

  • Priority 1
    • Display the current selection in the unexpanded box
    • Display a downwards arrow to the right of the box to indicate it has a dropdown
    • When clicking the box, or the arrow, expand to offer other selections, and when clicking one of those, contract the box and display the exact text for that selection in the box
  • Priority 2
    • Allow keyboard navigation – up and down should move amongst the items, highlighting as appropriate and return should select them as if the user had clicked them
    • Allow keyboard shortcuts – pressing a character should jump the selection to items beginning with that character
  • Priority 3
    • Spacial awareness – where there is no space for a dropdown to appear beneath, or it would appear outside of the viewport, it should appear above instead
    • Dimensional awareness – where there are too many items to display reasonably a scroll bar should display to keep the entire dropdown box in view

I have seen instances where many of these have been ignored, and in the very worst case, where selecting an item showed completely different text in the contracted box. Even an instance where the text displayed actually matched a completely different item in the list!

Hope this helps. AK