SEO Blog

Thoughts and tips from the expert minds of our team

02 Feb 2010

How to SEO Your Images

A couple of days ago there was a discussion amongst the the Bronco Team it started with Image results in Google for Kean Richmond when he noticed this picture for James ranking for his name :

Which was little odd, but if you look at the team page you see that James is after Kean in the pictures :

which got us thinking about the image before Kean which is Chris Antcliff and sure enough there was a picture of Kean in the Google serps :

the image files are named correctly and the alt tags are named correctly but when you look at the source code even though it’s formatted correctly we have it structured IMO incorrectly

Chris Image – Chris Image Title ( H3 ) – Chris Description
Kean Image – Kean Image Title ( H3 ) – Kean Description
James Image – James Image Title ( H3 ) – James Description

My gut feeling ( due to the fact I can’t be bothered to test and too busy to get Kean to restyle ) is that the correct flow of Image Optimisation in Google should be

Title Image Keyword ( h3)

Image keyword.png

Image Description

so what Google is doing is :

Chris Image Title ( H3 ) – Chris Description – Kean Image

Kean Image Title ( H3 ) – Kean Description – James Image

the correct images still get indexed correctly but the addition of the wrong images has to be down to the H tag passing proximity relevance to the wrong image and of course the more competitive the terms the less the H tag weighs.

Dave

DaveN

01 Feb 2010

WebmasterRadio.fm StrikePoint Show

It’s been a while since I have blogged about my Monday night radio show on WebmasterRadio show with Mikkel DeMib Svendsen. We do the radio show on a Monday evening from 8pm (Uk time) so tune in or download the podcast if you missed the show.

Here’s a taster of what Mikkel and I discussed tonight on StrikePoint

Mikkel started the show by explaining what has been happening in Denmark with Google Streetview. Last summer the Google street cars were out recording over 30,000km of roads in Denmark and once the data was released it has caused a lot of interest. There are big concerns as usual about privacy, but even more so in Denmark who have strict laws which prevents people taking pictures of people that can be identified in their own grounds or through windows.

Privacy is a big issue these days on the web but it seems Google has taken it to a new level, which can make people change their normal behaviour as a single snapshot does not tell the whole story. Even though you can report any issues to Google to get the image removed, Google isn’t recognising that some things can’t be undone and it would be too late. Getting some compensation from Google years later isn’t going to make things ok.

There are similar issues in the UK and the rest of Europe with Streetview, and do people really need the intrusion. An example of something really embarrassing is a Dutch fire truck being caught on camera running an old lady down … how do you explain that to your boss!

We then spent some time discussing the new Apple product launch. Mikkel first heard about the new Apple product from “The Onion” .. the Macbook Wheel! Unlike myself who is a bit of a techy geek and was waiting for the Ipad to be launched. But the Ipad doesn’t have a webcam, can’t access the Skype technology or read Flash. I reckon it’s a colourful Kindle without the bookstore. I don’t think I can be bothered though to buy the Ipad, and I’ll wait for the HP Slate. It also seems Steve Jobs has grown some balls or is crazy. Not to support Flash… crazy. Has he declared war on Adobe?

I talked a bit about the session I did at the London Affiliate Conference last week on the Demon SEO panel. I blogged earlier today about some of the topics that were discussed, and it was interesting to talk about these extreme techniques on a conference panel. Mikkel gave his view on this extreme black hat SEO and he was under the opinion that it’s best to be aware of what is out there, rather than putting them into practice.

Also today Distilled announced that they will open an office in the States as they partner up with SEOMoz and take over the consultancy side, so that SEOMoz can purely focus on Tools. So that will be interesting, but check out RavenTools as well as they have a good set of tools. We of course have our own set of bespoke super tools in-house which we use for clients, and one day may be released to the world. Mikkel mostly uses specific tools as well as some bespoke tools he uses in-house.

Then a quick mention of some conferences coming up including SES London in 2 weeks time. Both Mikkel & I are both speaking in London so it should be a good show. Talked a bit about some small gatherings / conferences I’ve been thinking about, and the plans for my book! It was interesting hearing Mikkel explain about his latest book and what needs to go into getting it right.

That was another show done, you can download it here, so tune in next week for the next edition of Strikepoint.

DaveN

01 Feb 2010

When SEO Gets Dirty

Disclaimer : I don’t condone this or part take in this, and this list is built up with 13 years of experience of what I have seen happen to sites on the web.

After the Demon SEO Panel at the London Affiliate Conference (LAC), (btw it wasn’t videoed so if you didn’t attend tough luck) I got to thinking about was it the right or wrong thing to do. I mean if you don’t warn people that it is going on then how can they implemented changes and monitoring systems, of course it also opened another world up to people that just want to hurt other peoples sites.

OK so what was discussed :

1) De-stabling the inhouse SEO team

Requesting references for certain members of the team from HR
Posting their CV’s on online
Posting their position on job boards
Employing them

2) Link Jacking

Buying hacked blog links
Adding them to known bad networks
Requesting link buys from Search Reps
Requesting links from bloggers that will out the link request
Buying links from malware sites

3) Hacking

Changing Robots.txt
Hidden Noindex/Nofollows
Add Large Cloaked Doorway pages then report them
Add links to their competitors
Webmaster tools account

4) XSS

Mess up their CMS
Gain Backlinks
Drop Cookies or add Affid

5) Traffic Flood

Destroy analytics data, so SEO team are busy deciphering it
Send Junk traffic such as adult etc
Ping of death
Slow ping of death

6) Paid Search

Clickbot their ads
Impression Fraud driving Adwords quality score down
If they are using inhouse tracking send on converting traffic to those urls
Test Inhouse tracking for Dupe content

7) Reputation Management

Trash them in forums
Trash them in Review sites
Mess with Google Suggests
Mess with Affiliates, overwrite Aff cookie with merchant cookie
Pay affiliates to bad mouth them
Rank scam sites behind them

OK, so what can you do to protect yourself? Having a solid in-house or external SEO team and good communication channels with the search engines is a very good start.

Some things you should be doing :

1.Server Monitoring – Stress test your servers, find out what traffic you can and can’t take
2. Stay up to date with sites like ha.ckers.org – news.ycombinator.com – seclists.org patch and stay ahead of the curve.
3. Watch your traffic, if you start to see unexpected traffic sources setup rules to bounce it so it doesn’t mess up logs
4. If you’re using off the shelf packages don’t add plugins without checking them and don’t let them auto update’
5. Take stock of your back links and monitor them, if you can’t account for IBL’s then check them out.
6. Push Good reviews and Newspaper articles to the top of the serps for brand terms, set up an alert based system to monitor your brand

Dave

DaveN

29 Jan 2010

Improving Page Load Time

It’s been a good couple of months since I was asked to write a post about reducing the page load times on this blog. In the meantime there’s been a big “will they/won’t they” around Google using page load times as a ranking metric, every man and his dog has blogged about the subject and every day Dave’s accusing stare is getting harder to ignore.

So, the blog is loading too slowly, apparently. What now?

Measuring Page Load Time

Before you can claim you’ve sped something up, you need to know how fast (or slow) it is in the first place. There are dozens of tools out there that purpose to give you that information. I’m not going to pretend to be an expert on all of them, and I’m sure others have written about them at length. Instead, I’ll just give you my main recommendations.

YSlow

Yahoo’s YSlow is an addon for an addon – the excellent Firebug for Firefox. It’s great. It shows an easy to understand and comprehensive report on how well optimised a page is for load time. Each of the criteria links to a page on Yahoo’s “Best Practices” site. It even gives you a straightforward A to F grade you can pat yourself on the back with after you’re done.

Pagetest

Not such a snappy name but extremely useful, Pagetest is an Internet Explorer based tool developed by AOL. I don’t recommend running it directly though, they have a publically accessible version with a few hosting locations around the world. Give it a try now on your own site. I used a number of other tools before finding this one and wasn’t too impressed – this one uses a real web browser to fetch the content so it’s a much more accurate measure of what a real user would experience.

The Rest

There’s also Google’s Page Speed which is similar to YSlow but I didn’t find it as useful. Plus it’s less pretty and isn’t that the important thing?.

Of course, Google Webmaster Tools now has a “Site performance” section under Labs. It’s presumably a good measure of the real world times users are actually seeing, but it doesn’t give much insight into the reasons for those times. I keep an eye on it but wouldn’t recommend it as your main tool – some of the advice it gives is downright infuriating. “Enable gzip compression for http://www.google-analytics.com/ga.js” repeated ten times. Thanks for that.

What’s Taking So Long?

If you’ve run a Pagetest report on your own site you can probably see it right in front of you.

In short: in short: external content being pulled into the page: images, JavaScript and CSS.

A web browser will have a limit to the number of simultaneous connections it will open to a remote server. If you have more images (or JS, CSS files) referenced on a page than that, some of them won’t even start loading until the first set have finished. You can see that quite clearly on this report. At the bottom of the ‘waterfall view’ in that report, notice that some of the images on the page haven’t even started loading until 7 or 8 seconds after requesting the page. Ouch.

In general, the fewer external files you’re referencing from your page the faster it will load. Unfortunately in the real world, you want lots of pretty images, CSS and dynamic content. So what can you do?

Speeding It Up

Rather than repeat all of the advice you can get from YSlow, I’m just going to list the big things that helped drop the page load time on this blog. In a sort-of descending order of impact:

Reduced the number of images on the page. I made as many of them as possible in to a CSS “sprite” – essentially, combining small images into one big one and using the CSS background-position property to show the correct part of the image. Take a look at the sprite for this blog and you’ll see what I mean. There’s a very quick and easy way to do this with a bookmarklet called SpriteMe.
Combined JS and CSS files: rather than one CSS file that imported three others, a big one with the content from all of them. Likewise for the Javascript content we control (not Google Analytics, TweetMeme, etc.)
Installed the W3 Total Cache plugin. I’ve yet to find anything better for speeding up the performance of a WordPress blog. It’s easy to set up and will cache pages and database queries, allows you to manage CSS and JavaScript files per page, automatically minifies the HTML, and more. It’s running behind some very high profile sites, too.
Set up some DNS aliases for static content – img.davidnaylor.co.uk, etc. The more hostnames your page requests from the more parallel downloads you get. Up to a limit. And each additional hostname adds an extra DNS query to look up its IP address, which adds up. This has been covered to death in much more detail than I could provide.
Moved this site to a dedicated server on a faster connection external to our office – at Fasthosts… which is whole other story in itself!
Made sure the web server was sending out compressed content (on Apache check out mod_deflate) to reduce the amount of data the user has to download, and that sensible “Expires” headers are being sent so caching will work (mod_expires)There are many other things to try that I haven’t even touched on here, but what I will say is that the law of diminishing returns is in effect, as always.

There are of course many other factors and optimisations to try that I haven’t covered here. These are just what I consider the low hanging fruit. If anyone has any easy one I’ve missed please leave a comment below!

Great Success

Of course, after I’d done all of that someone (who shall remain nameless) added a TweetMeme button to every post. That button pulls a JS file from tweetmeme.com that sometimes takes upwards of five seconds to load. A few of them on a page and my shiny new page load statistics go out of the window.

The moral of that story is that every time you embed someone else’s content into a page – be it YouTube, Google Analytics, Tweetmeme… – your total page load time is now at their mercy. Just something to keep in mind. That’s my excuse for not linking to a sub-2-second report at the end of this post.

My personal opinion is that, in the near future, I don’t think page load time is going to be a major ranking factor for search engines. From a user experience point of view though, it’s worth investing some time in.

James

27 Jan 2010

Twitter Local Trends

I’ve still got a hmm-ahh kind of relationship with Twitter. On the one hand it brings me content that I might not otherwise see, but as a communications medium I think it’s pretty woeful. Stupid little things like having to remember how many direct messages I had last time I checked so I can work out if I’ve got any new ones. Or the awful conversational threading.

And, btw, the exploit we mentioned last year – the one that made the pages of The Guardian and The Times? – still there.

Boobiewed??

But, today there’s nothing to blog about till Steve Jobs descends to Earth in a fiery chariot from Mount Ararat with his New! Revolutionary! device, so I suppose we should have a look at the latest Twitter feature – local trends.

Basically, you get to choose a country by means of… err… choosing a country in a little menu option and then you’ll see relevant Tweets.

Firstly, this is a pretty primitive way of going about things. While geolocating through IP might have it’s flaws it’s good enough for Google and pretty much everyone else so why Twitter are plumping for this “tell me what country in which you are located” business is anyone’s guess. Anyway, once you’ve done that, you start to see trending topics among other users from your chosen country.

The results are, predictably, weird. This morning, ‘mango juice‘ is a trending topic in the UK. Amusingly, almost every tweet about it read “WTF? Why is #mangojuice a trending UK topic?” which, of course, carries on perpetuating it. Maybe it will never end and will be a mystery for future historians to unravel, or the 2114 version of Dan Brown to base a novel on.

Actually, the answer is probably because not many people have set a UK preference. Probably 3 retweets is enough at the moment to set off a trending topic (as evidenced by the ‘boobiewed’ trend that’s cropped up since I started writing this post).

Slater also made the point that one of the great things about Twitter is that you get to see what’s happening around the world. Iran protests… plane crashes in New York… earthquakes… all these things are massive news and often surface first on Twitter, giving you a heads up on the news. Do you think Twitter is going to be enhanced if there’s another 9/11 and you’re busily trying to decipher why ‘mangojuice’ is trending?

So personally, I’m giving it a resounding ‘meh’. You, of course, will have your own opinion… and the floor is open as ever

paul carpenter

24 Jan 2010

Speaking at SES London 16th – 18th February

I will be returning to London shortly to speak at SES London which is being held at the Business Design Centre on the 16th – 18th February.

I will be speaking on the following sessions /panels.

Day 1 – Tuesday 16th February

12.45pm – PPC or SEO, The Ultimate Search Marketing Battle

2.00pm – Express Site Clinic in the Expo Hall

3.30pm – Industry Specific Search Strategies, Under the Hood

Day 3 – Thursday 18th February

11.30am – Tough Love, Get Your Site Tuned Up – Site Clinic

Look forward to seeing you there if you are attending.

DaveN

23 Jan 2010

Speaking at London Affiliate Conference LAC 2010

Next week I will be in London speaking at the London Affiliate Conference (LAC) which is the biggest iGaming affiliate event in the sector, and this year there are over 2000 attendees.

The conference is being held at the Old Billingsgate Market in London and runs from Thursday 28th January to Sunday 31st. The main conference days are Friday and Saturday and I will be speaking at 12.15pm on Friday 29th on the Demon SEO Panel joined by Bob Rains, Ziv Descalu and Frank Watson. We’ll be sharing lots of information on some of the more devilish SEO techniques out there.

Myself and Becky will be travelling down to London on Thursday afternoon and staying until the Saturday afternoon, so we’re looking forward to catching up with friends, associates and clients while we are there.

Next month I’ll be speaking at SES London, and I’ll soon be posting the sessions and panels I will be on.

DaveN

21 Jan 2010

SEO 101 Common Mistakes

For all the “SEO isn’t rocket science” crap you get from certain quarters it’s funny to see that companies from huge concerns down to one-man bands continue to commit the same errors they were making 10 years ago. If you’re an SEO, you could easily add to this list yourself (and I’ll have to thank the whole SEO team for chipping in about 63 ideas to bring this list to a nice round number!) If you’re a web designer who thinks that “good CSS = SEO”, a writer who thinks that “good content = SEO” or a developer who just thinks “SEO = bullshit” then here are a few pitfalls to bear in mind if you’re considering using SEO as a way to bring your products to market.

General Strategy

SEO isn’t just a discipline that exists outside the goals of your business. It should complement and be informed by wider business smarts.

Treating onsite SEO as a ‘one-off’ project without a plan to regularly review the site – especially if your site has a high product or content rollover, or has big seasonal changes to push new messages and offers
Changing horses mid-stream – revisiting keyword lists month by month in response to internal politics
Not consulting existing Analytics data to identify best performing keywords
Targeting all markets simultaneously
Forgetting about Bing and Yahoo, where rankings and traffic can be easier to find in the short term
Failing to understand (or convey to a client) that an SEO campaign is a long term strategy and results will not necessarily be evident in the first weeks or even months in competitive markets
Failing to utilize universal search options for increased SERPS visibility eg images, news, blog search, product feeds etc
Failing to work out initially if you can get a ROI from a sector you are targeting (profit margins, keyword volume etc)
Putting SEO in a silo outside core business objectives
Failing to include SEO input during the building of an online business plan and creating a site development spec.

Market Research

Concentrating on trying to concentrate on acquiring the ’same’ links as your competition
Looking solely at offline competitors
Not tracking industry news and events for new, fresh content ideas
Identifying ‘competitors’ purely based on results for broad, vanity keywords
Not using tools like Google Insight or paying for data from the likes Hitwise to identify seasonal trends
Not using the valuable data available from a concurrent PPC campaign to monitor converting keywords
Being unprepared to deal with social media
Failing to deal with negative feedback and reviews online
Failure to do your own market research through reviewing interaction with your site through Analytics, click tracking, customer surveys etc
Failing to have any form of conversion tracking software on the site to see what keywords are the ones that you have to go after

Keywords

Focussing on a small number of high volume ‘vanity’ terms rather than a deeper and better-converting long tail
Allowing keyword choices on the basis of “the MD checks this every day”
Chasing unrealistic keywords for your budget
Choosing keywords from internal industry-speak rather than consumer-led terms with actual traffic
Deploying brand / company name as part of a tedious “Company.co.uk – About” page title formula
Setting too many keywords to dilute linkbuilding and content efforts
Believing the numbers for likely traffic
Using the “other users found this page by..” method of including misspellings and synonyms
Forgetting that 25% of all searches have never been seen before and that search queries are typically much longer than single words
Not reviewing keyword choices to understand where your site is failing to convert visitors and why

Content

Copying content from other sites – potentially tripping penalties
Stuffing content with unnatural frequencies of keywords
Keyword “wishlists” in page titles (“UK SEO – SEO in the UK – UK SEO Agency from a UK SEO” etc)
Duplicated meta descriptions, which encourage Google to create their own snippets which can be nonsensical and harm clickthrough rates
Deploying content in images and Flash files
Creating content that has no value to human readers and fails to back up your market messages
Syndicating content to higher authority sites which are likely to be indexed before your own site and thus become canonical
Placing a large block of keyword-stuffed “seo content” a mouse scroll below the footer on the home page
Outsourcing content writing to the cheapest provider that you can find… you get what you pay for
Putting text within images rather than using background images under HTML text content

Watch your URLs

Not redirecting URLs to a canonical domain – leading to huge duplicate content issues
Leaving the non-www version and the www live simultaneously
Not sending correct 404 HTTP responses for broken pages
Using long strings of variables in URLs rather than short, static URLs with a proper file extension
Not using the correct 301 response for old content that has moved to a new URL
Using links for territories and currencies that create duplicates of your content in all but minor ways
Using ‘unfriendly’ characters in URLs, such as underscores instead of hypens
Allowing the indexing of URLS with session id variables
Not using keywords within URL structures over numbers and internal shorthand
Having a directory structure that includes terms like ’seo’

Channelling your Equity

Deploying sitewide links to low-value pages such as “categories” with 1 product in them
Linking every page to every other through an over-prescriptive menu and diluting equity spread to non-critical content
Leaking equity to external sites by not deploying the rel=nofollow attribute
Using ‘click here’ and ‘read more’ as default choices for internal links, rather than more descriptive phrases containing keywords where appropriate
Not using the homepage to channel power to the most important market sectors you’re targeting
Not using other properties you own (parent company websites, partners etc) to direct keyword equity to your target site
Using internal nofollows to try to sculpt PageRank
Failing to protect your site from exploits – everything from basic keyword spam in blog comments to sophisticated hacks
Using XML sitemaps to mask poor internal link structure
Not understanding the importance of ‘first link first’

Code

Deploying lots of inline Javascript and CSS and increasing the site’s download time
Keeping CSS and Javascript files on the same domain, reducing threading and increasing load times
Leaving dozens or hundreds of ‘keywords’ in the meta keyword
Having page titles that deploy “keyword wish lists”
Using navigation that can only be accessed through Javascript
Not considering the use of AJAX to bring in content and links to keep load times low and control equity spread without compromising user experience
Serving unoptimised images with large file sizes
Failing to label images with relevant alt attributes containing keywords as appropriate
Serving different pages to spiders and human visitors through cloaking without an obviously justifiable reason such as personalisation
Denying access to spiders through Robots.txt

Relationships

Not keeping the SEO company in the loop with changes to the company’s wider strategy
Allowing web developers to build/change things on the site willy-nilly without informing and consulting with SEO
Changing contact points frequently so that messages and learning get lost
Not introducing SEO agencies to other parties like offline marketing companies, PR agencies etc. This misses massive opportunities for content synergy and pooling of ideas.
Not responding to requests for information and content
Not ensuring that SEO recommendations are implemented as fully as possible
Blaming SEO partners for falling traffic without first seeing if there are wider market reasons such as seasonality that could be playing a part
Enacting SEO recommendations from other third parties without consulting with an existing SEO partner
Being unwilling to gain a small understanding of HTML / CSS
Not paying your SEO company!

The First Rule of SEO Club is…. “Don’t Talk About SEO Club”

Leaving “clues” in source code like <!– This content for SEO //–>
Using obvious file names and document structure. http://www.yoursite.com/styles/seo.css is going to attraction attention and all that “text-indent:-100em” stuff is going to highlight your hidden content pretty much off the bat.
Having dozens of obvious keyword landing pages linked from sitewides
Advertising the fact that you belong to a link exchange program by carrying banners that promote such schemes
Asking for advice about SEO issues on public forums without consulting your SEO company first
Leaving link requests in blog comments
Creating easily identifiable networks with common IP addresses, templates and outlink profiles that have an obvious relationship with your target site
Making sloppy link requests to bloggers who are likely to out you (hint: read their back catalogue!)
Using automated tools to check rankings on too big a scale
Using the same link sources for different target sites again and again