Peer Recommendation Is In Fashion, But There’s A Problem

With another article in last weekends Telegraph, with the Head of Bebo saying that peer recommendation is the way forward and the future no less of Social Networks, I can afford to feel very slightly smug that I’ve been barking up the right tree for the last 3 years, or more. Ms Burns says “I know and understand the power of search. However, social recommendations are the future…”

Kate Burns, Bebo Chief says peer recommendations are the future of social networking

My smugness should be short-lived however, as there are still two problems:

Firstly, institutional investors in Europe as yet seem unconvinced that now is the right time to invest in trust network technology or indeed associated consumer facing recommendation service optmised for mobile (Rummble has had lots of interest from VCs but thus far they’re yet to invest).  Business Angels have been very receptive, but the European VCs, in summary feel it is too early. We may yet move across the pond, to the sunny climbs of Silicon Valley, as a consequence.

Secondly and far more importantly, people (even the mainstream media, as the Telegraph article attests) are all talking about “peer recommendation” but forgetting a very important, indeed vital, factor: Trust.

There are a few problems with recommendations from your friends (i.e. your peers). Firstly, I have some fantastic friends whom I love dearly but with whom I do not share the same taste or opinions. That’s not a problem if I want to know from their newsfeed that they’ve gone on holiday, but it’s not so useful if I am recommended the Reiters Supreme Hotel they chose, in Austria, when I don’t golf and hate package holidays.  Secondly, even if ALL my 417 friends on Facebook, the 1400 contacts in my address book and the 500+ contacts on LinkedIn, ALL rated/recommended every place they went, there would still be 100’s of places in the world with no recommendations from them (and that is before you consider that they’d need to all use the same service or services between which recommendations were shared or collated).

While this does unquestionably serve to promote Rummbles trust platform technology as the answer to these problems, it is also the genuine flaw of peer recommendation services such as Geodilic and Whrrl, and of those services that blindly aggregate all data from your streams (e.g. FreindFeed).

Rummble Trust Network technology is not simply a social network, because it connects & infers trust (or lack of) between you and friends but also you and strangers

I am not saying the mentioned services are bad, I am saying simply that the focus needs to shift toward filtering information in a more intelligent way – reducing the noise.

Two of the Rummble team took 10 days out to build an experiment as a tentative step to help make sense of the burgeoning stream of data produced from Twitter; in reality the stream could have been from any social software. Tremors attempts to match Tweets with the venues they are from or referring to, to create a realtime feed of activity in venues across four major cities: London, New York and Austin Texas, with San Francisco to follow shortly.

Tremors is an experiment powered by Rummble

In this very basic first version, you can see what is going on at different venues in the city and it attempts to recognise some basic sentiment about the text. The next step, if we were to invest time in improving the accuracy, is that we could interpret the content as a recommendation (or otherwise) and ally this with content within the Rummble recommendation service.

Ultimately, we’ll be able to construct a trust network between users on Twitter whom may not even know each other, but share similar tastes or opinons, based upon an existing user behaviour. For now, it remains a fun experiment until we can resource improvements. You can read more about Tremors in the Tremors launch blog post.

Kate Burns of Bebo is probably right. Unsurprisingly, technology will evolve to build on and exaggerate existing human behaviour. We cherish personal recommendations because we usually have a basis on which to embrace or reject that recommendation. With social networks, I don’t always know how or whether to trust an acquaintance or a friend who may be outside of my core circle of trust, or whom I may respect but whom I may not know well enough to know if we share the same taste in restaurants or music.

In 2006 I presented a slide deck about “trust networking” and nobody seemed to understand what I was waffling on about; in 2009 I’ve presented a very similar a decks including trust networking and I often get nods of agreement and sometimes even yelps of excitement (see what I did there?

Our lives are built on trust and understanding (or lack thereof). Lets hope the Internet of the future is built on the trust and understanding part, not blind recommendations and lack of understanding.

Remember, “its about the data, stupid!”

Latitude: Googles Trojan Horse (or Why “Who’s Nearby?” Is Not A Business)

This post was published on Friday 6th February 2009 as a guest post on Mobile Industry Review

For the last 3 years now I’ve been crowing at conferences that “Who’s nearby” is not a business. I drew this conclusion from running playtxt, Europes first location-based mobile social network.

It started in 2002 and we had an Alpha launch in 2003. It was ridiculously early to market. Back in 2002 most normal people (i.e. those for whom a “tweet” today is still something only birds do) did not know what a social network was, let alone a mobile location-based social network. Thanks to MySpace, Facebook and the inevitable march of technology, even my own mother is now aware of social networking, SMS and GPS.

By 2005 Google had bought our main competitor Dodgeball and although the mobile operators were still charging for Cell ID lookups (ludicrously, they are STILL trying to!) I already believed it was only a matter of time before location became a commodity. It would too easy to do for start-ups to do and even easier for others such as Facebook, which was on its ascent.

I decided that “who’s nearby?” was never going to create a multi-million pound business and I made three predictions, some which are still relevant today:

• GPS will be in every phone as cameras were then becoming. (GPS chipsets are extremely cheap, power consumption is becoming lower, processing power higher and Galileo is on the horizon -literally, haha).
• One of the gorillas (Google, Yahoo et al) will release a free Cell ID/Location API. (Google have and its excellent).
• “Who’s nearby” will also become a commodity

For the last 2 years I’ve been telling any start-up which is building its own Cell ID database, that it must be mad. I see no business model. Google about as likely to charge for Cell ID lookup as it is for its maps API; and that likelihood is slim.

There was (and is) money to be made with tracking and Cell ID technology, but both industries begin with “S” and neither spell the world “Social”. Instead, it is Security (child tracking, staff tracking) and of course Sex (proximity dating, adult services); infact any vertical where a premium can be demanded – we know that fear and shagging both command strong emotions which can result in a buying decision. Wondering “Where are my friends?” does not; unless of course you’re intensely paranoid or have VERY accommodating friends.

There is no mobile internet: there is only the internet.

This has been my other crusade for the last 2 years; and this is probably what Google believes too. What I mean is, that fix-line world-wide-web access is the black & white TV of the internet. Amazing in itself, but without the full functionality of what we recognise as “television” today.

Location, portability and the need for personalisation (a mobile being such a small, personal device) are the three missing dwarfs which give us our Seven Dwarfs of the modern internet. (The first four were IMHO: the web browser as user interface, freedom to publish without government or minority corporate control, always-on fixed cost access, and broadband bandwidth; Snow White being the internet itself).

So in the near future (3-5years?) no one will talk of the “mobile” internet but simply, the internet. You will have an iphonesque device (in size & looks if not in O/S which you take home and plug into your 24 inch screen and keyboard …we’ve still a decade to go before we type goodbye to Mr Qwerty and say hello to HAL.

Be under no delusion, Latitude is Googles Trojan horse into the social networking space.

After Googles purchase of Dodgeball it was clear they had every intention to roll out a service such as Latitude and they are perfectly positioned to do so.

Almost by-passing online social networking entirely (aside from Orkut which only took hold in Brazil) I believe Google will pursue a wide-reaching mobile social play. Google will build up a critical mass of users on Latitude; and they will join because:

• It is Google (so its trustworthy; yes still)
• Its easy to use – simple UI and simple privacy model: Automatic, Manual or Hide your location (or as I prefer: Honest, Lie or Paranoia)
• It has reach (27 countries at launch, lots of handsets, no GPS required)
• Its free

They will then likely launch an API (in the process solving some of the standardisation and interconnectivity problems – possibly using the new OAuth hybrid or equivalent) but also roll out other functionality enhancements. Although the latter may take longer than you think.

Latitude has lots wrong with it too e.g. Gmail import only (where is XFN Social Graph import or device address book comparison?) status update is crying out for Twitter integration and a hook into FireEagle (with which Latitude does not compete, yet) would all be very welcome (the last two are unlikely for political reasons but would be a fantastic nod to the open ecosystem) and dont forget part of Latitudes beauty is its simplicity; and Google have time on their side.

Many of us have been waiting for location-based services to come of age for YEARS! but in reality we’re still in the early adopter curve. Infact, I’d go even further than that. At BeingDigital in 2008 I stated on stage to a deluge of ridicule, that Social Networking wasn’t yet main stream. The laughing continued until I asked how many parents AND siblings of delegates had email? The answer was predictable: virtually everyone. Then I asked how many parents and siblings were also on a social network; over 75% of the hands dropped.

150 million people on Facebook is a lot, but 3.2 billion people have mobile phones: that’s a lot more. Email is mainstream, social networking is still maturing. Eventually it will of course become part of everything we do “online” rather than be a destination, with your social graph becoming portable and also actually owned by you, not FaceSpace.

So what does this all mean?

1) Location is already commodity AND your friends location will become a commodity.

Any service will be able to plug in and use this data (with the right permissions). Its already happening – checkout Yahoo’s FireEagle which is an aggregator of location between services.

2) If you’re a start-up building LBS, Cell ID, friends nearby services, or anything else which is being commoditised as we speak, see above.

Loopt; west coast startup run by a bright 24 year old entrepreneur – nice guy, flawed business plan. $13million+ in funding, nudging just 1 million users after 3 years with low engagement metrics. Differentiator? There isn’t one. Case closed, game over.

3) If you’re running anything with the words “mobile social network” in the title, lock yourself in a room with your team and work out how you’re going to save your business.

That means innovate. Mobile is not a differentiator, its an inevitability.

At Le Web 07 I met with Christian Wiklund, Founder of Skout. He had built a cool location based mobile social network (LoMoSoSo anyone?). By Q1 2008 when I met him in San Francisco, he’d already realised that competition was fierce and the concept was flawed — and that was before the gorillas had waded in. I implored him to change strategy (something which infact he’d already started doing). He chose dating. It’s a smart move. Dating generates money—and lots of it. Proximity dating, or infact “mobile dating” in general has never been done really well (even Mr Arrington agrees).

As a LBS start-up, you need to think about adding distinctive value for users; differentiating on location is an oxymoron. I know some of you are making money, some of the pure play mobile social networks are even profitable – great. But there’s an iceberg ahead and it may be bigger than it looks: just ask Captain Edward John Smith.

The future is relevance; the context of not only where I am but what I’m doing, who I am, where I will be. In summary: It’s about the data, stupid.

..and that will be what I write about in my next post; if they’ll have me back!

Why are mobile social networks not as popular as their online cousins?

I recently rote a post to the MomoLondon in reply to Paul Walsh from Segala’s post about Mobile Social Networks; and thought I’d share my reply here:

Rummble is a service which incorporates all aspects of a mobile social network which you mention; although to describe it as a social network would be inaccurate as that is not our core proposition.

Having starting building Europe’s first mobile social network (playtxt back in 2002) we soon realised that any big consumer “mobile” social network would have to offer something more pervasive than just location based messaging and “who’s nearby” functionality, in order to become valuable. That is unless you serve a specific vertical (e.g. Mark Curtis’ Flirtomatic which is doing very well).

Rummble focuses on making it easy to find things nearby that you will like – i.e. location based content from within your extended social network. However, by nature of knowing the users location, we also tell you which of your friends are nearby, which are going to be and also the ability to message those who are nearby you don’t know, without revealing your or their mobile number.

Dopplr is trying to build an entire service around saying whether your friends are going to be in a location or not; where as we believe this is a natural function of something more pervasive and have offered this functionality since our playtxt platform.

What is changing now, is the ubiquity of location information and user uptake of it.

We believe GPS will become the next defacto addition to mobiles (as the camera was) and in anycase there are other ways to find location without paying the operators for it . The iphone uses SkyHooks technology (wifi positioning and I believe also cellID) and there are other options out there (Googles MyLocation is another good example which 70% of the time locates me perfectly – despite the log-in disclaimer stating “This service is reliable to within 4000 metres”

Rummble supports GPS, we’ll support Yahoo’s Fire Eagle API by next week, we’ll soon have an iphone app out and we’ll support Google’s Android when it arrives. (btw, if people don’t know Tom Coates‘ FireEagle then check it out – its an extremely important step toward distributing your location reliably to services and will raise awareness of LBS outside of geekville. They recently launched Fire Eagle at E-tech but there is already a fervent and growing developer community).

You do raise a very valid point about having a decent and simple UI. This is actually our current battle and having built a bullet-proof backend its the GUI we’re going through iterations on to get right. As all developers on this list will know, its an annoying and costly process to get even your mobile XHTML looking good on multiple devices, let alone a Java app with its archaic high-level restrictions or having to write native apps for Symbian and soon Android to get something looking good.

There is a reason why its taken until the woolly world of Web 2.0 and cheap hi-res TFTs for us to get easy, attractive and people friendly web applications and interfaces on the desktop – Browsers are pretty much consistent (within reason) and broadband enables UI lovelies such as AJAX for asynchronous saves without page reloads and lots of colourful gradients (too many some might say!).

We’re a world away from all that with mobile. And there in lies the problem. We can produce a wonderful UI for the iphone -but that’s a drop in the ocean. The resources and time it takes to guarantee are fantastic mobile user experience are too high for many start-ups. And the user does expect a slick easy experience because that’s what they are used to on the broadband web.

Can you imagine the step-change even if simply all phone browsers supported AJAX; even on EDGE/GPRS the lack of need for an XHTML page load would revolutionise the user experience.

At Plugg in Brussels yesterday Rudy de Waele sited Buzzd (which won some momo awards) as a popular mobile social site – but its interface is pretty practical and for good reason.

We think Rummble provides pervasive functionality – the challenge for us is to make that usable for the majority of users who dont have an iphone or the equivalent shiny new HTC.

The cost to start-ups of providing slick UI across so many screen sizes is frustrating for us and gives users a nasty surprise; but ultimately we’ve got time for these things to change, as until MNO’s bundle flat rate data with every single tariff they offer, we wont see the hockey stick curve of mobile internet take up by Joe Bloggs what we all crave.

Person Rudy de Wael

Right click for SmartMenu shortcuts

Thoughts on Fire Eagle from Yahoo

In Fire Eagles inaugural TechCrunch review, the line that jumped out was the terribly misguided leader “Fire Eagle launches with near zero functionality”. Were Fire Eagle a traditional location based service, this would likely be fair; but surely TC is missing the point. FE sits at the apex of services, providing a consistent route and API so that whichever consumer service a users uses, they can consequently propagate that update (at their discretion and convenience) to all the other services. In a nutshell this can be done:

* Directly at the FE website / direct via SMS etc
* via an FE compatible app
* or via a future FE social app

The 3rd option is my concern; if only because users need to be very clear where FE primarily sits as a service. I’d even say there was a potential argument against FE producing a (e.g.Facebook) app because it should be the FE application developers who develop the apps and services which then utilise FE as the location propagator between them. This then retains value for the app developers in having their users interface directly with them. In this respect TC seemed to miss the point and failed to focus heavily enough on why IMHO FE is an important catalyst for all of us working in LBS.

This leads me to a final point, which is are there any plans to support signup of users to FE via the API (I haven’t checked so if it does apologies) in parallel and at the time of signup, e.g. when they signup to Rummble they become a member of FE by default? Likewise combining this with FE Support for OpenID would further accelerate FE’s chance of survival and acceptance as a defacto location propagation tool, and hopefully ensure users understand FE’s remit as primarily being the glue between services not a service itself (assuming of course that last point is true!

Oh, and whoever said to TC that Fire Eagle is like “Twitter with location” should be shot — its a bad comparison for all sorts of reasons, not least because Blaine & co are still struggling to keep their services reliable, but mostly because it completely undersells the effect that an industry wide propagation service for LBS may have.

If the woman sitting next to me on this flight to SXSW doesn’t stop making deafening suction sounds chewing her boiled sweet … there will be blood.

Andrew.

Fire Eagle open location API Launches (E-Tech 2008)

E-Tech, San Diego – The first major development in a while that may help the Location Based Services out there realize their true potential. Sorry the video is poor quality; and the sound may be pretty dodgy too!