Archive for the ‘Search’ Category

Tweeting about places not X/Y coordinates

March 3, 2010

In a recent conversation with a fellow tweeter I was asked why I on earth I was spamming the world with 4square tweets rather than just geo-enabling my tweets. I think this completely misses the boat. People don’t care about your X/Y coordinates, unless it’s a way of you telling them what’s near to them. Machine’s care about X/Ys and rightly so. People about places and that’s exactly what makes 4square of whatever mobile social local app du jour you favour work so well.

It’s nice then to read of the recent announcement that twitter is going to start allowing people to structure conversations around places, rather than geo-coordinates.

It also means that Twitter is noticing what’s going on on top of their API. Some of the most popular clients are leveraging the latest location-aware devices, be they Androids, iPhones or soon always-on sat nav devices to let people tell their friends (and noticably, their enemies and everyone in between) where they are at. Or to cut a long story short: Twitter is going local. Just like everybody else.

Technorati Tags: , ,


Google’s IYP killer

October 6, 2009

In Google creates a new simplified ad unit for local business, Greg Sterling takes a new advert for small to medium enterprise, i.e. the Internet Yellow Pages market, through its paces. Google has been trying to enter the SME market for a while, but from everything we’ve seen their churn has been horrific. This means despite the numerous incentives that Google throws at SMEs to get them to sign up to start Adwords — normally anything from €25-€200 free ads — most SMEs quit within their first year. Google has hinted for a while that it will start a simplified approach that will appeal to small business owners, but has not done much in the market apart from some opening moves in the Local Business Centre.

The classical trifecta are all there: a great landing page, transparent reporting and the ability to buy additional visitors and leads. What makes this particularly interesting (and confrontational to IYPs) is that SMEs can buy traffic and leads based on a ‘cat-locale’ i.e. a combination of location (business or search) and search categories. In addition, typical IYP product add-ons such as a click-to-call are included, except now of course the call ‘whisper’ (the brief introduction at the start of each call before it’s connected) says “This call is brought to you by Google” and the delivered lead is reported in your overview. As you’d expect, reporting also means you see where your clicks and leads are coming from. Pricing is flat rate and contracts can be cancelled quickly.

Google continues to test whether (as IYP wisdom says) you require a real sales force to get SME market penetration. It is interesting that Google’s basic problem is less in sales and more in retention — i.e. customer acquisition seems adequate but their churn is painful. IYPs have started investing, and in a number of cases, succeeding with fully automated solutions to deliver guaranteed click and lead products (common in Europe, but for how long, we shall see) and budget-based packages. If they can do it without customer care, why can’t Google? It seems more likely that Google’s go-to-market and product approach was wrong. Whether they can achieve the penetration they want only with online sign-up, or whether they’ll need to take on a larger cross-media advertising approach, perhaps combined with T-sales, remains to be seen. One positive side effect of Google’s approach up until now (at least for Google) is that although SMEs may not have stuck with Google they are unlikely to have gone back to traditional IYP products since the perception is radically altered once you’ve bought a few clicks at €0.50 each, i.e. Google SME products accelerate the market fragmentation, even if they are not themselves successful. In this way, they become part of the baying pack of dogs trying to bring IYPs down, even if they don’t become the new top dog.

Although this is only a test in a couple of US metros, I can’t believe that IYPs will take this lying down. Google is still pretty dependent on IYP and telco feeds in the US, so I’d expect US IYPs to kick them back where it hurts: this means war.

How to move from print to online

February 12, 2009

Techcrunch informs us that Google has recently a Finnish paper mill and intends to build out a data centre there instead.

Maybe that’s something Yellow Pages companies around the world should look at too? Can’t imagine the business case is going to be quite as dramatic just yet…

Technorati Tags: , , ,

Zcapes is live

February 11, 2009

It looks like the first phase of the Zcapes go live has started. You can check out (in your mobile browser but your desktop would work too) and start reading some of the content and widgets that are already there. You’ll have to wait another week or so before you can create your own Zcapes, but that’s not so far away.

When I first wrote about Zcapes, I was in I love the idea that context is a new long tail and the use of different plug-and-play meta-data and functional services and bots to define it. Since then I have talked to a couple of people who are also trying to get at the basic building blocks — the genotype — of what makes an online experience interesting and who are trying to find how to capture these basic elements into a consistent whole. Zcapes looks like a fun way of playing around with some of these ideas.
There you’ll see a range of services and bots, e.g. messages, poll, twitter search, flickr search, RSS feed, RSVP, tagging & location. Although you can’t create any yet, and the widgets are reasonably limited, you can browse some zcapes through their tags. One to follow.

Powerful white pages when it doesn’t brown out

January 29, 2009

TechCrunch has a great article on, comparing them to and They explain that Pipl does better (‘scarily’ better) than some of these competitors by crawling the ‘deep web’. This means that Pipl spiders enter commonly used terms in search pages and interact with other dynamic pages in a more intelligent way. This means that they find pages that most spiders cannot access from static crawl directories and sitemaps and in this way extract more information. This is not as sophisticated as it sounds. Many crawlers use this approach to generate new seed pages to crawl from. I think the ‘deep web’ is more marketing than technology. Nonetheless, it does well.

I did a test search for my dad, Michael Allenson in the UK. It’s very useful that you can specify a country and that Pipl will only use sources for that country. The site found two traces of my father from (a scientific information search engine, which I helped build a l-o-n-g time ago. At that time it ran on FAST, just like the Yellow and White Directories that Truvo runs) — two patents that he authored in 2000.

Most searches for Allenson on search engines, including people searches, return information about Gary Allenson, a baseball player. By being able to exclude US information, Pipl did better than most from the start.

A useful feature is that Pipl returns a number of image results inline with text links (a bit like This meant that when I searched for my friend Nuno Macedo, although it quickly found his LinkedIn profile I was able to select this from the results even more quickly by clicking on his picture.

Sesam, a Norwegian search engine has for some time allowed users to search for a term, then show related images, web results, businesses and people that it’s extracted from the search results. This means that if you search for Steve Ballmer, both Bill Gates and John Markus Lervik (CEO of FAST, which was taken over by Microsoft) will come up as associated people, you’ll see some pictures of Steve (including one with his tongue out).

When I last visited I got the following front page message:”

Sorry, we’re very popular…

Due to an unusually large amount of traffic we had to disable the search access for a short while. Please be patient while we upgrade our capacity or bookmark our site and visit us later.

Thank you for visiting Pipl.

Re-tweeting as a rank for twitter

January 12, 2009

Twitter allows you to re-tweet microblog posts you like to your followers. If sufficient people re-tweet, positive lock-in kicks in and the tweet takes on a life of its own. This can mean that checking re-tweets becomes a powerful mechanism to rank ideas flowing through the twittersphere. I really enjoyed Re-Twit’d which shows the top re-tweets.

The nearest competitor I can think of is the wonderful Twitscoop which shows a continually evolving tag cloud of the twitgeist — “what’s hot on twitter right now”. I also like the key tags appearing in Google’s universal search results.

I’m going to playing with Re-Twit’d in the next few days. It’s great to see more and more apps that are not owned by twitter, but use the API to make more sense of what’s going on in the twittersphere. I like the eco-system that is being build up around basic app.

Zcapes — contextual miniblogs

January 12, 2009

Zcapes is a new Dutch start-up. I love the idea that context is a new long tail and the use of different plug-and-play meta-data and functional services and bots to define it. Meta-data covers

  • the obvious — you, who, what, where and when
  • the emerging — physiological inputs like heartrate, accelerometers
  • or combinations like the geolocation of your social graph,
  • deductions: if you’re moving on a train line, you are on a train; if you’re at a friend’s house with many friends, you are at a party

Zcapes helps segment information into appropriate contexts so that you can find the useful and usable information quickly when you’re in the context. You could think of this as the first stage of implicit search. Since who you are, what you’re interested in and who your friends are are part of your context, a first step in segmentation is individual personalisation. The only way to get that to scale is to enable people to publish information services that are useful to them, and allow others in similar contexts to find it more easily. Zcape promises to allow people to interact and transact in specific contexts, aggregate the relevant content for them. SPRXMobile, the company behind Zcapes, believes that this will make advertising something between push and pull. This is a rather familiar refrain, but one you can imagine working if the contexts are sufficiently granular and there is sufficient content. It has failed pretty spectacularly pretty regularly mostly because contexts don’t work or content is lacking (e.g. Facebook advertising), but it is unobtrusive and successful in a few situations (e.g. Adwords on Gmail).

One nice thing that drops out of this context platform is that many of the common business ideas can be recast as zcapes, e.g. geotagging iPhone or Flickr photos or playing certain music for different segments of your regular run, and yet it also gives some ways to add more social aspects to this, e.g. seeing who plays what music for location, seeing what twinsumers based on context buy in the store you’re in now. I think a key question there will be sufficient context granularity and functionality that users will want to build contextual microblogs in zcapes rather than a geo-aware twitter with photos (like twinkle).

Some of the services and bots SPRXMobile have pre-announced include the pretty ubiquitous weather for your location, flickr pix, microblogging with twitter and facebook and RSVP for invitees. It look like the launch use case will be: meeting up with a few mates downtown night, who’s in and (via implicit context search) where shall we go and what shall we do?

Indeed some of the example zcapes include a regular Friday movie night at work. Later on the roadmap are things like what is currently on TV.

All-in-all it looks like a promising mobile context platform. I for one will be trying it out when it releases

Turning away free business

September 28, 2008

One of the things I love about the iPhone is the integrated Google maps application. It’s great for local search. Although the content is pretty poor in the Netherlands compared to the Gouden Gids, it’s readily available on the phone and once you do find something, it’s easy to add the contact into your phone with name, address and telephone number all in there.

Recently, I was looking for a fridge for three months, so I thought I’d call up Boels Verhuur. These guys rent out almost anything. Although they focus on the B2B market, they do a lot of business for consumers too. So I search for Boels Verhuur and get a few hits near my house. I call up the first one and he says “Boels in Leiderdorp. Ohhhh – you’re probably looked for us in Google, right? No, this is Boels in blahblahblah, you need to call this number”. I call that number, and I get the answer “Yes, we’re Boels in Leiderdorp, but we don’t rent out fridges, you need our affiliate Party Business”. Well, to cut a long story short, it takes five different phone calls and two mistaken website searches before I find the affiliate which rents out fridges. I reckon 5% of the business leads that come in this way will get through the funnel before they even look at the products and services available.

It has become a small and medium business imperative to have a single place that lists correct uptodate contact information alongside the most important brand, product and service information. To ensure that it is really uptodate and correct, the control of the information needs to be in the hands of the only people who really know that: the business itself. The business will automatically be ranked more highly because the information is unique and fresh. Ideally, this page will also list out aggregated opinions on the business and opinions on the products and services that the business delivers (which is what people will often be searching for), and do so in a way that allows the growing volume of people searching semantically to find the reviews and contact details directly.

Perhaps this will also be the page that paid search traffic is directed to and for which the business gets regular and clear reporting. But these are secondary to having a single place for your uptodate business information that is easily found organically.

It’s a small investment to get lots of free business leads. It’s also a small investment to avoid the Boels Verhuur experience above. It was particularly frustrating that the business itself knows and recognises that it has an issue — probably from the volume of leads coming in to the wrong place — but has not taken the simple steps needed to fix it.

Technorati Tags: , , ,

Arbitraging Microsoft

September 10, 2008

Great post in Think-Through on an arbitrage scheme using Microsoft’s recently introduced cashback approach.


I love the very idea of arbitrage — ‘risk-free money’. Anywhere else, this would be called a scam and the would be overtones of theft. But in the world of finance, where it is understood that markets are (mostly) efficient and that there can be no such thing as a free lunch (mostly), nobody seems to mind if you happen to find a lunch that’s free because they are safe in the knowledge that it won’t be free for long. Interestingly, some of the thought leaders in capital markets theory do describe a number of free lunches. But because they know that free lunches are not possible, they describe it as reward for extra risk taken.


I spent a year or two working with a treasury derivatives trader building treasury derivative arbitrage models. These were humungous Excel sheets loaded to the gills with add-in functions, some custom C code, and pulling in realtime data, detecting small differences in pricing between synthetic and real securities. An example: if you take a euro-swiss franc swap and a swiss franc-dollar swap you can create synthetic euro-dollar swap. This has (just about) all the characteristics of a real euro-dollar swap but it might have slightly different cost. Where the difference exists, it’ll be tiny but if you have a billion euros on your books overnight, it may be something you can take advantage of ‘for free’. Most of this simple type of arbitrage really does not exist in the markets any more, but when we used more complex treasury derivatives, and worked out all the permutations across a basket of ten currencies, there were more opportunities.


I still remain interested in financial arbitrage, and on the lookout for simpler ways of consistently making money — but that’s for another post.

Google’s AdManager comes out of beta to form a real challenge to OpenX

August 26, 2008

If it’s good enough for Google Suggest — Google has recently announced that they their AdManager has moved from private beta to a public release. AdManager is aimed at publishers with small direct sales teams. It includes some nice features out of the box, mostly focussed on tracking directly sold and network-based inventory, then enabling the sale, measurement and provisioning of this. Naturally, there is tight integration with AdSense and multi-lingual support is excellent. There is also reasonable third party network integration. Some of the time in beta testing has led to new features such as time dependent rollout and previewing of ads. A sophisticated inventory management and provisoning system like this is no small investment and it’s a classic Google move to make it free to level the playing field — or to raise the competitive moat to monopoly level, depending on how you look at it. In this case, existing competition gives away its software to advertisers, and makes money, just like Google does, from the advertisers.

It seems unlikely for competition like UK-based OpenAds (now called OpenX) to be running scared. With chairing by former AOL head Jonathan Miller, more than $20 million in backing and an impressive roster of customer names, OpenX may be just looking to take their 30,000 customers into the arms of a competitor. However, Microsoft got hold of Atlas when it acquired aQuantive, so why it would want OpenX is anyone’s guess.

Note that AdManager is fully hosted and a closed proprietary system, unlike OpenX which allows its publishers to host it and to extend its PHP code base. OpenX’s CTO, Scott Switzer has also pointed out that a good number of publishers may not want to add another piece in the Google monopoly puzzle. On the flip side, many small publishers may love the ease of how  software-as-a-service enables them to get live quickly.