Log in

No account? Create an account

Mon, May. 7th, 2007, 02:12 pm
All the. Small things.

Way back, when I was at Yahoo!, there was naturally a lot of discussion about the meteoric rise of Google. At the time we'd dumped Altavista and were using Google (which Yahoo! owned a largish stake in) for the web crawled results. As a first cut at stopping the user exodus we switched to using crawled results primarily rather than the directory results we'd previously used. We blended directory results with Google's results.

This did stem the tide but one of the interesting things for me was talking to people about why they used Google over Yahoo!.

Now, to me, it was obvious. For various reasons (mostly to do with the layout of the page and the way that adverts were inserted) we were much slower to load and the page was much more cluttered. But when asked about why peoples' response was that Google's results were better. But how could that be? We used Google's results. It was all about perception and it was the first time I realised that there were going to be times that I couldn't code something better it had to be all about impressions. This in turn caused a large amount of my money to be transferred directly from my wallet into Jeff Bezos' as I started buying more and more books on usability and design. I never got any good at it, mind you, but at least I could get the general gist of what was going on.

In retrospect this may seem obvious but at the time it was a big revelation to me. And also sort of a relief - it meant I (or my ability to program) wasn't wholly responsible for everything.

Google's stuff was so good because they made little incremental changes. Sure PageRank was a better ranking algorithm but what people really liked was the clean, fast interface. Ditto Apple. Ditto Flickr. Stuff that seems obvious in retrospect but of course, nobody had done it till then. Like all the best things.

Take Google Maps. Personally I don't like bits of it and I prefer the normal A-Z style maps of Streetmap when I'm trying to find my way around. But the really nifty thing is the ability to scroll round the map but just dragging with your mouse. So, so obvious in retrospect but no-one else had done it.

Shortly after Paul Rademacher did a mashup of Google Maps and Craig's List called HousingMaps which overlaid properties for rent or sale onto a map. Genius!

Meanwhile the rental market in London is completely moribund. Before the web really took off everyone used a daily magazine called Loot which had listings of stuff for rent. You turned to the section that represented the area you wanted to live in, scanned for properties that looked ok, looked up the address in the A-Z, worked out how long it was going to take you to get to work via public transport (or Shank's Pony if you were lucky enough), lather, rinse, repeat.

When the web came along all that happened was that those listings moved to the web. The only real advance was aggregation companies like Vebra that scraped the results from multiple estate agents (realtors in the US) so that you only had one web site to trawl through.

But it was still a pain in the arse.

Foxtons managed to make their own site much better than the crowd with a clean layout and easy to access to maps, photos and floor plans. Of course they had access to much better meta data but still, it was a huge leap from what was available before.

Then came Nestoria. Taking the HousingMaps concept one step further they aggregate a bunch of sources, extract meta data and then plot the results on an inline map and then, to really spice it up, they blend in local council information such as nearby schools and hospitals, plot the congestion charge zone, show tube lines and, more importantly they show good pubs using data from Beer In the Evening. Then they make their urls RESTful and easy to construct and sling RSS in with everything. Want to have an RSS feed for any 2 bedrooms flat between £500 and £750 a month? No problems. Any time one pops up it'll turn up in your RSS reader.

"So what?", you might say "Isn't that all really obvious?". Well sure. But why isn't anyone else doing it? Why is Nestoria so much better than any other site out there that I've seen? I mean sure, there are technical difficulties doing this sort of stuff - read Marc Tobias Metten's talk on Geocoding that he's given at the London Perl Workshop and other venues for some of the intricacies or Nestoria's own blog for more details on that sort of stuff - but really it's inexplicable.

Now, full revelation time - I know the web team behind Foxtons and also the crew at Nestoria. Indeed a couple of the Nestoria peeps worked with me at Yahoo! during those interesting times so you could accuse me of bias or shilling. And indeed, I'm inordinately proud of them - they clearly learnt their lessons from that time even if I (or senior management at the time) didn't. But I think my point is very valid - it's not often you come across something that's just obviously better than the competition in every way- for example I think Flickr's great but I'm pretty sure my Dad doesn't want to use the social networking side of it. But the last one I can think of before Nestoria was Epicurious whose Faceted Classification of courses, ingredients and cusines was just so obviously right it was blinding.

What's this got to do with my usual search ramblings? Nothing much really - it's a wet, rainy bank holiday monday and I'm mildly hungover and just pondering stuff whilst catching up on old episodes of Sports Night and thinking about a conversation I had at London.pm on Thursday.

Mon, May. 7th, 2007 04:06 pm (UTC)


Good grief. That's like a middle-class North Londoner's wet dream come true.

Tue, May. 8th, 2007 02:26 am (UTC)

It's very shiny

Tue, May. 8th, 2007 08:24 pm (UTC)

I think at least part of the reason people said that Yahoo's results sucked is because they were *used* to them sucking and so were more inclined to be critical of the shiny new Google results, if they bothered looking at all.

Wed, May. 9th, 2007 07:35 pm (UTC)

stef magdalinsky likes to say you can fix a bad back end with a good ui, but you can never fix a bad ui with a good backend. i paraphrase, but it's about like that.

even ui people often have a hard time understanding the roll of perception in user evaluation. jared spool did some fantastic work there: he idetified for instance that people perceived amazon as the fastest e-commerce site, when in fact it was the slowest. turned out that the perception of speed was tied to the user's expectation of making a successful purchase, rather than actual speed, or even making a successful purchase. i think google's success is kind of a slippery thing. for many years after google had 'won' v yahoo, yahoo in fact had more users, and they spent more time on the site. but at the same time it seemed no one used yahoo- everyone anecdotally was hemorrhaging to google. enter an explanation via aaron marcus, a very good ui guy who applied the research of geert hofstede on culture. google's interface was clean, and fast, - and risky. there wasn't much to tell you how to proceed. google didn't hand hold. yahoo was amenable to helping the less bold user. the page set up yahoo's authority and options gave naive users and risk averse users context and a way to get into the web slowly. yahoo stuck around with the user for a while, google got out of the way. that means google was very good for people drawn to high risk professions- say banking, wallstreet, tech, and journalism. yahoo was better for people in most other demographics, people that didn't want to range out onto the scary net on their own.

eventually, though, being more likely to be used by journalist meant google became the word for finding things, not yahoo. also, i've always thought that yahoo was full of people that would have demographically speaking rather have used google, and that eroded their value as the paternalistic portal. (i mean that not at all insultingly) i wonder how much the fact there there are few hand holding searches these days means that many people aren't getting as much out of the net as they used to, but what's done is done. the next generation won't need to anymore than bicycle training wheels.