Last night, I was searching for wellies for my sister’s Christmas present. Despite all the grumbling I do about Google, it was still the place I began my search (although I eventually ended up back on eBay.)
The first page of results had one organic search result above the fold on a 1366 x 768 monitor resolution. And that was Amazon, who were also taking multiple paid slots. In fact, increase the screen height to 1000 pixels and there was still only 1 organic result, with no fewer than 18 paid slots (if you include Google Shopping).
Congratulations to Amazon there for taking the sole organic slot (as well as 4 of the paid slots).
Now, I am not suggesting that every Google result looks like this, but treatments like this for commercial niches are becoming increasingly commonplace, and there is no compelling reason for Google to roll them back. Commercially speaking, filling the page with paid ads is simple logic: after all, Google gets nothing from the sites it shows for free. From a technological perspective too, why plunge billions into fighting spam and ranking an index when you go can whitelist a bunch of sites for most of the big money terms and leave the long tail to random chance?
Seen in this context, domain clustering is no mere mistake. That ‘used ford focus’ today returns a mere 11 domains for the first 3 pages of results makes eminent sense. Google are simply saying to everyone: “you can’t compete in this space unless we think you’re a brand and have whitelisted you as such”.
Of course, when you’ve arrived at that point you’re a whisker away from being a paid directory with an ancillary web search.
Ironically, of course, Google have long proclaimed war on paid directories and have been busily purging them from the rankings for years. Equally ironic is their war on sites that they judge to be “too commercial” by carrying many banner ads or even Google’s own AdWords ads.
For SEO as an industry and a standalone activity, that leaves us expending our resources chasing an ever-decreasing pool of eyeballs. It’s not that you can’t still succeed in driving traffic through organic channels – you can and always will be able to – but we have to be close to the endgame of the idea that you can pitch into a truly competitive space with SEO as your central/only strategy.
While SEOs will continue to do the best they can to make sites friendly to Google and (hopefully) admit defeat and turn off the blog networks and comment spam robots, it doesn’t leave much for anyone to actually do. “Link building” is a philosophical dead end – a mere adjunct to PR, banner advertising and partnership building either through commercial relationships or social outreach.
People will still report success for their efforts and doubtless correlations will continue to be drawn between link building and sporadic SEO success, but if you’re truly looking to the long term: Google isn’t the sandpit it used to be.
Tuesday’s random thought: once you’ve worked up close and personal with Google for a while – particularly if you’re interested in trying to frig the rankings in your favour – you start to really appreciate the flaw with Google as a model in itself.
Firstly, it was built as a scientific tool: hence its historic reliance on links. Scientific papers reference each other as citations and thus the link became a proxy for citation in Google. A scientific paper that is referenced a lot is clearly important and thus, theoretically, so is a website or domain.
As we all know, that model fell down the second that people realised that Google rankings were monetisable. Hence, we are in year 14 of the link building wars.
But, browsing eBay for Christmas shopping at the weekend really brought home to me flawed Google is in other ways.
- eBay stock is categorised by its sellers when it is added to the site. Thus, there is no need for them to try and reverse-engineer a gimcrack categorisation system after the fact and try to force advertisers to enact some abstruse, incomplete technical standard.
- eBay search can be organised in several useful ways that Google doesn’t particularly have an analogue for:
- Cheapest product first – self explanatory
- Cheapest product including delivery – equally self explanatory
- Distance to seller
- ‘Buy it now’ prices
- Time remaining on auctions
- Most reliable/top-rated sellers
Combining your search term with these things make it a relative breeze to find the right thing on eBay. Better yet, it’s all done within one interface. If you click 10 results on Google, you will find yourself faced with 10 different sites, with different logos, navigation systems etc etc. We all know that even for relatively sophisticated web users it can be a nightmare finding your way around an unfamiliar site.
eBay is also truly a level playing field because there is no algorithm trying to assign weight to all this stuff – it’s just some relatively simple database querying. If you are the cheapest, you can be found by those motivated by price. If you are the best in terms of customer service, you can be found by those who want the security of reviews. If you are selling locally then someone nearby can find you.
By trying to build algorithmic proxies for these things into what was originally built as a citation engine, Google has gotten just plain bad at an awful lot of stuff because it is trying to imply structure on a web that is fundamentally unstructured.
I’ve averred before that I would never now go back to Google to book a hotel now that I’ve got a great app to hand (although I might use it for ancillary information about the location of the hotel in relation to nearby attractions). It’s also fairly obvious that you might use Google as a conduit to finding insurance prices, but that the actual action will ultimately happen on a price comparison engine and once they’ve captured your details, will you really go back to Google to start again next time?
Should you build a website to sell online?
My sister in law was asking me recently about getting a website built for her to sell her handmade jewellery. I told her not to bother, because the cost of having a site built would be one thing, but actually getting it turn up in Google and thus generate business would be another story entirely.
An eBay store would effectively cost her a fraction of the price – most of which would only be paid after the fact on actual sales. She could also compete fairly on reputation and price and not through buying into some bullshit link building scheme that she wouldn’t even understand (and shouldn’t have to) and that could burn her business to the ground at some point even if it was successful in the medium term.
More seriously and pertinently to my own interests, if you search on Google for used cars of various makes, you’ll increasingly see the likes of Friday Ad and Gumtree cropping up. With the best of intentions, unless you’re actually know something about cars, the very last place you should buy one would be from some Tom Noddy on Friday Ad or Gumtree. Yet – and on the basis of domain strength or brand or somesuch nonsense, Google are happy to serve up those sites up.
I’m possibly being behind the times here, as I don’t keep bang up to date with all PPC developments any more, but this definitely interested me. Don’t ask me why I was searching for a bouzouki.
I’ve seen image ads before – particularly for eBay – and have dabbled in them myself by pulling images and individual stock items into the ads through from a Google Products feed, but this is clearly something entirely different (shopping results are also in the results, but inserted midway through the regular organic positions, in the same way that local/news results are).
This is an entire block of ads with imagery and product details pulled into the ads without going through any intermediary such as Google Products. It’s a very interesting model and I’d love to see it rolled out to the car sector, where my interests currently lie!
Naturally, it’s a further distraction from the organic results. As I mentioned above, Google Product results are also injected into the middle of the regular SERPs meaning that with this ad block and the other Google properties on display, there’s very little space for the SEO-led Bouzouki retailer in the UK (now that’s a niche!)
Anyway, no time to give this any further thought at the moment, but if you see anything similar cropping up I’d be interested to hear about it.
Firstly, these ads appear in between Google Shopping results and the Gooooooogle pagination thing. Obviously that placement isn’t accidental: to get to the next page of results you are now forced to see the ads. That can only have a positive impact on CTRs.
Secondly, anyone who has used product links or sitelinks in their AdWords know what a dramatic effect they can have on CTRs too (I personally saw a fivefold increase in CTR for ads with sitelinks on a big enough sample size to know it was no fluke). What I haven’t seen – and I could be behind the times here – is the promotion of product links with no ‘plus’ button: the images are just planted straight into the ad.
It’s a very attractive format for an advertiser and one that would certainly appeal to me – especially if it were rolled out into the main SERPs.
Does that take Google another inch to being primarily a shopping and advertising channel to the further subtle detriment of the organic SERPs? Well, duh. Everything Google does at the moment is informed by that rationale.
I know that sites need advertising to survive. I also know that overly intrusive advertising can kill the user experience. Check out this visual nightmare from Metro.co.uk that I was presented with when trying to read about the fat sleeping benefits supervisor that Lucien Freud made a star – a full 850x1025px of graphical madness. “Less is more” is clearly not a memo that reached the boys at Metro / O2.
“Both Google and Facebook offer their services to users apparently for free. But the services are not of course free. Both companies sell the information that users provide – in search data, or personal profiles.
Apparently, these companies mostly sell their (your) information to advertisers who mine the data in order to target consumers more effectively. But, despite fervent declarations about transparency, in fact, it’s very hard to find out exactly to whom they sell the data or what the “data miners” do with it. McDonald’s or the CIA? We’re not told, even though it is information about us that they are trading.”
So sayeth the Guardian. Given their political stance, it’s little wonder that they immediately leap from ‘advertisers’ in the generic to “McDonald’s or the CIA” in the specific. In the gothic imaginings of the left, The CIA and McDonald’s are handy ciphers for Government and Capitalist evil respectively – which pretty much informs the tone of the whole post.
In fact, neither Facebook nor Google sell your information to anyone. Instead, they allow advertisers to target you through information freely offered by you. And it’s all done at an aggregated level. In, well, the same way that the Guardian itself operates.
“We also ask some further, voluntary questions so we can gain a clearer understanding of our users. Your responses help us to sell appropriate advertising space and so keep the site free. They also enable us to personalise services for our users. We do not share this information with third parties unless you have specifically consented to this.”
No-one at Glaxo Smithkline is sitting there reading specifically about me. Or you. Or anyone. They have, however, instructed Facebook to show an ad for washing powder to any male between 30 and 45 who talks about trousers. Just as Google sells search slots based on, well, searches.
But hey – if you’d rather do it the old way and just get blanketed with formless, unpersonalised mailshots – don’t use Google or Facebook!
If you’re a small online player trying to punch above your weight with traditional PR, 2011 is the year to start questioning that particular route to market. Firstly it’s been an interesting and revealing few months for the meeja as we’ve come to know it. As its once-mighty institutions are revealed to be hollow shells, even the perceived value of ‘news’ coverage has to be in decline.
Secondly, for web-based businesses who are in a position to directly measure the impact on their site (rather than just nod their way through a collection of press-cuttings) the impact of dead tree coverage of web businesses is shown to be pretty poor.
An 8 page colour pull-out in the Sunday Times might set phones ringing, but will deliver next to no impact on site traffic without amazing optimisation. And this isn’t speculation: I have seen it with my own eyes. Beyond a vague and contentious blip in daily direct visits and type in traffic for the brand there was zero effect on the site overall. No lasting uplift in traffic or conversions through the site. And, of course, until technology improves a little, there’s no way to capture SEO juice from dead tree coverage.
An isolated example? Here’s some screenshots of what a 2 page spread in a Sun pullout (daily readership: 7,5920,000!) did for traffic on the day it was published.
The branding was featured extensively in the Sun coverage – but there was no discernible activity in brand search activity.
Domain Type-in Searches
OK. Clearly a discernible uptick here – the brand was mentioned primarily as the domain name in the coverage. But look at the scale. Type-ins went from a normal daily average of 10 per day to a whopping 20 on the day of publication.
Perhaps a small increase in people coming directly to the site – but the number is at best equal with other random spikes during the preceding few weeks.
Coverage was strictly analogue. The story didn’t even reach the digital version of the Sun, so there was neither traffic nor linkjuice to be had. Perhaps had the story reached that level of penetration into the paper’s coverage this blog post would have a different tone.
And in the meantime…
I set up an AdWords campaign to tackle the same theme. This was for two reasons:
- To pick up residual traffic from people who’d seen the story on the train and gone to research it on their PC at work/wherever.
- To widen the coverage to tangential subjects and themes that didn’t appear in the press story.
It was a very limited affair – 13 keywords and 2 ads and a daily budget of just £50. Despite that, it is now clearly outperforming the Sun in terms of response.
The brand has been seen by far fewer people (the hundreds, rather than the millions) but of those hundreds, there has been a much more immediate and direct involvement with the brand.
Measuring a couple of events in Analytics has demonstrated beyond doubt that people are interacting with the site in a highly targeted fashion through this channel in a way that seems highly unlikely that The Sun delivered.
A small note
If any PR companies are reading this, I’d just like to make it clear that it was a great PR scoop for us – a new, small business – to get that kind of coverage. There’s certainly no doubting the effort and skill that went into getting us there, nor the positive tone of the coverage.
There is no criticism implicit in this post of the PR activity itself, just an airing of doubts about whether the trad PR model is the best fit for a companies who operate mainly in the SEM/SEO spaces.
PR works on the principle of brand exposure. See the brand a few times in reasonable context and you have a brand presence which you can then exploit to sell your wares more easily.
Once you are a brand, PR becomes a natural part of everything you do: your every move is reportable and interesting to the press and public and you’ve probably got budget to finesse the coverage to suit your needs.
But, as a cost, PR is a long term activity. Monthly retainers are part of a picture that then expands to include junkets, offline stunts and the inevitable costs of managing things, brainstorming ideas and so on.
But for a lot of web businesses – particularly start-ups – that model just doesn’t wash when you look at cold hard figures.
If money is tight, don’t necessarily become entranced by your picture in the paper and consider spending the money on PPC.