Tim Nash UK SEO Blog

Latest Comments

Sounds like you have been busy on another great research project. After seeing Google display my Adsense id and email accounts on search pages on another browser tab , I quickly realized that I was being tracked.

I had little doubt that Google tracks everything we do, they’ve not exactly made a secret of it. But, in my limited knowledge of cookies and tracking it seemed quite probable that others could gain information to my accounts if I surfed while logged in. Your findings prove my suspicions to be correct.

Good for marketing, I suppose; but horrific for personal data security. The last 4 digits and type of credit card??? Guess, I won’t be using Google Check out if I can help it.

Reply | Original | Permalink | Share | Tweet

That only helps on session exit not while your currently surfing though it will help with the css history aspect.

Reply | Original | Permalink | Share | Tweet

CCleaner rules!

Reply | Original | Permalink | Share | Tweet

At 2 in the morning it was an impressive feat!

Reply | Original | Permalink | Share | Tweet

intricacies of binary extraction great line lol, Dr Who Rules !!

Reply | Original | Permalink | Share | Tweet

aw, sorry mate. never meant to insult you. au contraire, it should be seen as a compliment. all the best mate. read you next year.

Reply | Original | Permalink | Share | Tweet

That’s quite insulting but I will dump the analogy and keep the younger part

Reply | Original | Permalink | Share | Tweet

Wish i could make it tim, but i’m still stuck in sunny SA, surfing and jet skiing and doing all the other things that makes for a hard life

someone’s got to do it, since everyone else is having fun at conferences and what not.

merry christmas to you and yours mate. and respect for being honest and open; a younger (not nearly as grumpy) version of michael martinez.

Reply | Original | Permalink | Share | Tweet

yep I will just fetch the sword

Reply | Original | Permalink | Share | Tweet

yep you’re going to have to take the blame

Reply | Original | Permalink | Share | Tweet

ah crap, I would blame wordpress but since I have been using it for a couple of years now that might not wash…

Reply | Original | Permalink | Share | Tweet

at least I can link to Al correctly

DaveN

Reply | Original | Permalink | Share | Tweet

I probably owe you a pint or two!

Reply | Original | Permalink | Share | Tweet

Hi Tim,

Guess I’ll get to eventually meet you in person, have just booked my ticket, looking forward too it already.

Have a happy holday,

All the best for 2009 as well

Paul

Reply | Original | Permalink | Share | Tweet

Hey Tim

I always believe in the saying; “Success is in the details”. I’ve been playing with the locations of the backlinks on a website, and did several tests myself.

My initial test results have shown me that the location of the backlink on the page is an important variable when the overall value is calculated by search engines (especially google).

But when you get a backlink, overall, the goal is to have a pure organic backlink (at least in google’s eyes). What is the main goal of this backlink? To have a higher CTR, OR a higher backlink value for search engines.

For me the business is all about ROI, and in most cases backlinks with less search engine backlink values have brought me higher ROI.

So, the value of a backlink should be divided into 2 categories; SEO value, and the business value. Of course we are talking about initial values here. ‘Cause we all know, a higher SEO value will also increase the ROI in the long term through organic search results.

I don’t want to change the topic here. Just wanted to say, how great of you to share this valuable study with us.

Thanks again.

Ahmet KIRTOK

Reply | Original | Permalink | Share | Tweet

Using GAs is normally one of those things people start of thinking about them but then quickly decide to use weighted probability, but they were perfect for our example as we had a continuous feedback loop so while our initial dataset was small as it grew and as we started introducing some additional factors (such as had they visited our clients competitors) we could really start to make use of these as mutators. If we had simply used weighted probability for assigning we would have had to manually access new factors and have spent most of our time trying to get the balance right.

Reply | Original | Permalink | Share | Tweet

Great post Tim. It’s interesting to see a real-world application of genetic algorithms returning tangible results. I’d love to hear more - were you involved in that part of the project too?

Reply | Original | Permalink | Share | Tweet

@ Andy From memory the PPC and organic traffic were up by 4 to 8% while the Affiliate traffic was up between 5-9% on the landing pages when averaged across the winning recipes in each experiment. Its worth remembering that is an increase from pages that were already statistically performing well so while they may not seem like a huge amount it made a large difference to the bottom line.

The big thing was the introduction of profiling across the board which literally paid for itself within an hour and is now at the core of their business interaction with users which is not bad given how sceptical they were initially to the idea.

@Ahmet Your affiliate company will know where your redirect is regardless from their traffic logs so they always have somewhere to start from depending on budgets and how much detective work they want to put in they can normally find the original source of the traffic, while their are plenty of ways to conceal your traffic sources, to be honest its very rare a company will use their affiliates traffic sources though they may feel the need to protect their brand for companies which rely on affiliates it makes no sense to upset the hands that feed you.

If your interested in how we track and monitor as well as weight sites you might find this post on Link Worth interesting.

Reply | Original | Permalink | Share | Tweet

Tim, thank you for taking your time and writing this great, long blog post. Very informative post.

Do you mind sharing little bit more information about how you’ve found the 122 info domains of the affiliate?

I know many affiliate marketers try several ways to hide their source of traffic. Because we hear horror stories of advertisers stealing the profitable source and keyword data and going direct every day.

Some more technical information to affiliate marketers about hiding their sources would be greatly appreciated.

I’m also all about A/B testing. I have been talking about the importance of A/B testing and it’s funny to see even some top ecommerce companies are still not good at data mining, and A/B testing.

Looking forward to seeing more case studies.

Thanks again.

Ahmet KIRTOK

Reply | Original | Permalink | Share | Tweet

Tim, great piece.

You forgot to wave your flag a little, Marketing Experiments style

% improvements in conversion?

It might be a little sensitive, but you didn’t name the client.

For many this would seem like a lot of work, so an indication of the benefit to be realised would be appreciated.

Reply | Original | Permalink | Share | Tweet

I spelled ‘hear’ wrong - how embarrassing -

Reply | Original | Permalink | Share | Tweet

I’d be fascinated to here more from you about this: “utilising users own searches and introducing behaviour mechanics to identify user looking to compare/browser vs purchase.”

I wrote a brief blog post for a client recently about that very topic…essentially it was about ‘converting keywords’…keywords as both a call-to-action and for semantically driven SEO - inferring the different forms of searches and designating the right landing page. Good stuff - looking forward to it!

Thanks again Tim,

Ben

Reply | Original | Permalink | Share | Tweet

Worth noting I said “our normal product profiling” that doesn’t of course mean we disregard them within our profiling but if left without some marker or with profile groups related to a product category they could easily upset the results.

I’m glad you enjoyed it and hope there was the occasional tidbit, segmenting a user group should not replace contact or personalisation its nothing more then a pointer system. The best systems rely on the purchasers profile, previous purchases, current or recent searches and what ever else they can get their hands on.

One of the things I didn’t get into which I might do in a separate post since this is in theory an SEO blog is utilising users own searches and introducing behaviour mechanics to identify user looking to compare/browser vs purchase.

Reply | Original | Permalink | Share | Tweet

Hi Tim, just finished! What a piece - really great work!

Before I was an SEO/SEM I was more of a traditional marketer - loving my stats and analysis, digging-up information where ever I could…so this article really struck a couple of great chords!

One thing that did occur to me though from reading this piece is that the power of mining and grouping your target is massively important, but it’s important to recognise where the market could be expanded too.

For instance, when you say:

“During the purchasing process its a great idea to understand the reason for purchase this simple easy to answer question allows us to discard the user within our normal product profiling…”

…it’s important to remember that even if they are not part of their profiling, a friend or member to their family might be, so some exposure to segment-specific special offers (for instance) might go a long way. Especially at this time of year when people are shopping for others on mass…

You hit the nail on the head, when you said:

“You cannot know your customers individually and eventually you have to accept that you are going to have to start grouping them.”

Grouping goes so far, but I’d suggest you (anyone) doesn’t over specify their segment needs as you could alienate some. Maybe low of little of wiggle room?!

Awesome post, plenty of food for though and really well written. I’m becoming more and more a fan of your work every time I visit!

Cheers,

Ben

Reply | Original | Permalink | Share | Tweet

I hope you get to the end I of course started it as a quick case study but thought it was missing any practical tips so added bits to help others (maybe)

Reply | Original | Permalink | Share | Tweet