Omniture's mini summit 08 in London

Everyone else is doing blog posts about their experiences of going to conferences (see Kate Duffy's descriptions of what went on at Search Engine Strategies in New York and James Kelway's description of what he enjoyed at the Information Architecture Summit on his Userpathways blog). Therefore I think I should probably write a brief post about my experience at the Omniture Summit in London last week. In reality it was only a brief summit, but I think some of what I am about to describe to you may strike a cord.

On Monday afternoon whilst half of the building were doing some nice Omniture University sittings on Search Centre and SiteCatalyst the few of us in London who are on HBX came together for a little discussion on HBX's future.

For those of you unaware - HBX has been bought out by its biggest competitor Omniture. Omniture have said that they are going to continue to support HBX for as long as clients are still using it, but they won't develop the platform. In fact, they are going to take everything that was good about HBX (the report builder for example) and replicate it in SiteCatalyst. Then they are going to try and persuade you to move to SiteCatalyst by giving you lots of really nice migration help.

Migration help seems to be a good idea in principle - everyone going on to one tool helps SiteCatalyst and in theory SiteCatalyst is better than HBX. The help that Omniture provide will get you up and running on SiteCatalyst with most if not all of the functionality that you had on HBX. Given that they are going to tell you how to tag your pages, help you come up with a naming convention and presumably show you how to use the tool, that should make this migration extremely attractive for lots of websites.

Anyway, moving swiftly on from the debates about how best to move to SiteCatalyst and on with the proper stuff on the Tuesday:

Josh James is very enthusiastic. I'd go and watch him speak again. Does he make me want to join his company - I'm not so sure. It's interesting though that both he and Shar Vanboskirk both talked about how we weren't fulfilling the true potential of our online Marketing in their keynote speeches. However that was only partially of interest to me (along with the interlude by Josh about the football game he went to see in Prague), the second set of sessions was a bit more of interest.

I started the second session by going to see the first session on 'Evolving from Data Manager to Actionable Strategist' with Adrian James. Unfortunately his analogy and gags comparing it to the farm that he grew up on in Scotland went down a bit like a lead balloon and I felt a bit sorry for him (given that you have to explain what a burn was). And actually most of what he said whilst being true, is what we should be doing anyway:

His suggestion is that by setting up alerts, predetermined dashboards and pushing reports out automatically (demonstrated of course by using SiteCatalyst) it should free up more time to be able to do proper analysis of the sites. Whilst this is true, I am always wary about pushing out these reports because you end up having to do more work to explain them - my preferred stance is to send them out with commentary so that actually the person on the other end doesn't really need to read the report and could just read your analysis.

Anyway, moving on I next went to see 'Driving Revenues by Using Data To Power Your Email Marketing' which is actually something that I've looked at in the past with previous companies. Ted Wham who is the SVP of Epsilon who look after our emails here was a really good speaker (although he darted round some things I was interested in). Apart from getting us all interested he showed us a nice little calculator that you could use to prove the return on investment. Unfortunately it was the return on investment for an eCommerce site trying to get users who've fallen out of the process to return and buy using email analytics. Very clever, but not that relevant to me at the moment.

Then I had the 'Increase Your Effectiveness Using Vertical Industry Metrics & Best Practices' for the media sector with firstly Matt Belkin and then a bit of a case study by Shorful Islam who is the BI Manager at ITV Broadband. This I thought was the most beneficial of the presentations of the day. This was the one that was actually aimed at people like me.

Matt started by talking about Fusion and how you could use SiteCatalyst to introduce more metrics into your analytics solution. Which metrics? Well if you can upload ad impression, cost and sell through rate data that you can then map against pages, you can convincingly create a page impression vs ad impression report. This can then be taken further to introduce a revenue per page report. This can then be taken a step further and introduce a revenue per visit report and so on. Ultimately what it should lead you to do is work out which pages are the most profitable and optimise your strategy for driving visitors there. He then pushed on to how you measure your video to ensure that the advertiser's details are shown in the video (hence increase the volume of sponsorship).

Then we moved back up to the 'Social Networking and Communities' session with Martyn Jobber. This was very interesting given that I have been blogging about Social Bookmarking recently. Although Martyn (quite rightly given the audience apart from me) was more keen on promoting the method of measuring those people as opposed to growing them. It's interesting though how you measure the effect of your social networking (probably beyond me in this post).

Finally we had 'Advanced SEM Optimization [sic] Techniques' with Giuseppe Sessa. This was interesting in that he showed us some of the more advanced things that you can do with SearchCenter to optimise your campaigns. Again, largely irrelevant if you don't have SearchCenter because he didn't really go into any of the detail about how you come up with the ways of showing that you need to optimise outside SearchCenter.

Breaking into Social Bookmarking

This post has been a bit of time coming, I think. Actually the last time I wrote about Social Bookmarking, I had been thinking about it for a while. It's not a particularly easy topic to write about and I always end up waffling, so you'll have to bear with me tonight. What I want to talk about is my experiences with social bookmarking on this blog and in general. But we'll start from this blog:

www.stumbleupon.com

I've been playing around with Stumbleupon for a bit now. It is interesting for me because for the sites that I look after it provides a steady stream of traffic. It is the most consistent of the lot of them - it generates small volumes of traffic in a constant stream as opposed to the large spikes that I see on other social networking sites.

How does it work?

You download the toolbar for your browser. You browse the web. When you find something you like, you click thumbs up in the toolbar. If nobody has seen it before, then you can write a little blurb about it. You also have a button that says 'stumble'. You click on it and it takes you to a random page that other users have thought was good. If you set up your profile correctly it takes you only to the pages that are in your interest realm. You can view lots of interesting pages and just stumble onto a new one. It's quite addictive.

What happened to whencanistop.com?

Whilst playing around, both myself and Henry decided that we would put a thumbs up to the whencanistop home page. You can give it a thumbs up too if you like. In fact, if you do like a site, why not give it a thumbs up - it'll give other people who like what you like more chance of finding it. Having got two people to click it, I suddenly got a spike in my site traffic:

There you go - I got 118 visits in basically the space of an hour. Did anything come of it? Not really - it was fairly untargetted and I'm a real niche website. But most websites aren't - you should really be able to find people who will be interested in your subject line.

What should I do?

Get everyone you know to download the stumbleupon toolbar. Get them to put thumbs up to pages they like and thumbs down to ones that they don't. Include your own pages in this analysis. It's a community - so if a page is good then it'll get lots of thumbs up and do well. If it's not good then it'll get thumbs down and will drop like a stone, probably dragging the rest of the pages on stumbleupon with it. This means that there is no point spamming by giving thumbs up to all posts.

http://www.slashdot.org/

Slashdot was something that I joined a couple of months ago and had been looking around, but hadn't really been doing anything with until recently. Until the other week that is:

How does it work?

Slashdot works in a slightly different way. Firstly you don't actually have to be a member to post things, but it helps (I think). What you do is go to the slashdot entry form and fill in the url that you want to post as well as a little information about it. Slashdot's very efficient moderators then go through the entries and decide which ones are worth putting on the site. The stories that are posted stay on the home page until they are shunted down from above. Every day Slashdot sends its users an email with the stories from the previous day with links.

What happened to whencanistop.com?

This is where it all gets interesting. I was looking at a post by Aurélie Pols on the OX2 Web Analytics blog about Yahoo! taking over IndexTools and making it free. I gave my name as my blog post and submitted it. Imagine my surprise not long later when I noticed that I was top of Slashdot!!! Anyway, I wondered where the traffic would go for this one and it clearly wasn't to me:

I'd be interested to hear (if anyone over OX2 can tell me that'd be great) how much traffic that it generated for the other links in the site. The thing I did notice though was that for days afterwards I was getting referrers that were clearly email programmes - people clicking through on the email that was sent afterwards.

More importantly though - look what it has done for the number of links that I have to my site now on Yahoo! Site Explorer (this may look much lower now, but it was up in the 60s at one point because of all the RSS feeds). Add into that if you do a search in Google for whencanistop, whereas before there were very few pages, there are now hundreds (801 for me today) - even if I now have been pushed off number one by slashdot.

What should I do?

Sign up to Slashdot. Submit articles that you think are interesting. Don't spam. This is really important, so I am going to say it again. Don't spam. Seriously - only 10 or so articles go up on their pages every day. If you keep sending them submissions that don't help you, then you aren't going to get any response after a while because they'll recognise you and filter you out. Only submit stuff that they'd find interesting - it's mainly tech and science.

www.digg.com

I think I mentioned this in my last report. Digg is an unusually site because it seems to generate huge volumes of traffic (or can do if it sits for a long time) but doesn't seem to have any obvious pattern about when (even if this person claims to know)

How does it work?

Digg works in yet another slightly different way (although it is more similar to Stumbleupon). You have stuff on your site that people like. They digg you. How do they digg you? You see pages with little 'Digg this' signs next to them (stories on the Guardian always seem to have this). I'd recommend it. Sign up to Digg and then digg your favourite stuff. Digg has a system whereby if it gets a certain number of diggs in a space of time, it will shoot it up the home page and then get more people going to it. If there are lots of people going to it and digging it again, then it will get even higher. If people go there and don't digg it, then it'll slide back down again (very similar in this sense to www.reddit.com).

I'd recommend signing up to all of these services and using them frequently. How frequently? All the time. If you're looking at the bbc, digg it. If you are researching an article or post and find something relevant, stumble it. The thing I like about stumbleupon is that it lets you look at all the pages you stumbled. You can then use this information to help you write your story (because all the links that you need are in one convenient place).

And don't forget to stumble this blog :)

Yahoo buys IndexTools to Continue Consolidation

Seeing as I had the temerity to link to Aurélie Pols' blog post about Yahoo's acquisition and subsequent decision to release for free on Slashdot, I thought that I'd better write a post on it. Not least because I had a presentation today from Kate telling me that everything was about scent (ie making sure that you consider where your users come from to give them the best information on the page). What would be good on my home page when I have a load of people coming through from an a slashdot post on Yahoo buying IndexTools than a post on Yahoo buying IndexTools.

Whilst I could go on about how this is good for the public, Yahoo and IndexTools, I think I'll take a slightly different approach and look at how it is going to impact Businesses and the web analytics industry in that order.

Firstly if you are an IndexTools client, what must you be thinking? You've just paid a fortune for this new tool (actually, it might be really cheap, I'm guessing, but it'll probably seem like one) and then they've announced that they are going to make it free. Well actually you need to read this carefully. They are only going to make 80% of the tool free, the other stuff - the clever stuff that gives you the deep insight, that is going to still be paid for.

If you were thinking of being an IndexTools client, then this is great. It is great for two reasons. Why bother with an RFP when you can just go and get 80% of it for free, test it out then work out if you want the rest of it. You can even work out if the amount of money that you are going to pay would be worth that extra 20% or whether actually most of what you want comes from the first 80%. It depends on how much you rely on your stats for your business decisions.

If you were debating whether to pay money for a tool or just to get something free, then this is quite good as well. You can put GA or Microsoft Adcenter Analytics on your site as well as IndexTools. Then you can tell people that if they give you £x then you'll be able to help them answer y question that they have. And it'll be easy because you'll have tagged your site (I'm hoping that the full version and the shareware have the same tags).

If you are going to just leave your site on a free tool then this is great too. I have my blog on Google Analytics and Microsoft AdCenter Analytics. I do this because not because I am an analyst (ok, partially because I am), but also because they give me different things. I'm hoping that Yahoo! won't just attempt to copy Google Analytics and will try and take a different stance. There is no reason why you can't have them all running in tandem.

If you are with an enterprise solution (and I liked Aurélie's response to Omniture's response to the announcement of the acquisition), it will be interesting to see what happens in the long run. Eric Peterson thinks that it may not just be 80%, but that many clients are going to get it all for free. Does this give the people who are contemplating migrating from HBX to SiteCatalyst something else to think about? Well yes it does because now there is another option. Stick with HBX until you can get a free IndexTools account, then simply decide whether to roll up to the enterprise version when your HBX contract runs out is now an option.

Omniture will claim that they don't really go toe to toe with IndexTools. They may be right. The sites that use SiteCatalyst and HBX are probably bigger than just SMBs and the tools certainly have a greater functionality than Google Analytics or Gatineau at present. However, previously the differences in prices was between a small amount and a not quite as small amount. Now the differences is free and a not quite so small amount. Free things always seem better in the long run. Omniture really need to get a wriggle on in combining those best of qualities between SiteCatalyst and HBX to make a super product that nobody could do without.

All of it really helps bloggers. We don't spend money on adwords or ppc campaigns (generally - I don't make any money out of this site, so why should I spend anything other than time on it) therefore we don't want to spend any money on analytics. Essentially we may just get better at working out what goes on with our traffic and how to build better sites and write better content because of it.

Microsoft Ad Center Analytics (aka Gatineau) Reviewed

Because I said I would two weeks back, I thought that I'd give you an update on what I thought of Microsoft's Gatineau (or Ad center Analytics as it is now called). These are just my initial thoughts, by the way, so I am hoping that I can give you a more in depth analysis in a couple more weeks time when I've got more data and I can do some proper comparison to it and Google Analytics (effectively its main competitor).

My first comment is that I am British. In this country we spell it centre. I don't mind other people spelling it center - it may be the common spelling in the country that Microsoft comes from. Gatineau on the other hand is spelt the same in English as it is in American (as it is in French, Spanish, etc). Just because you've migrated it into Ad Center, doesn't mean that it has to be called that.

My second comment is around the request that I join up with Ad Center in the first place. Now I understood that this was to be on its way out, but it appears not. After that it is really simple to do, although I am still of the opinion that what you actually need is real step by step instructions here. A lot people won't understand the automatic ftp that you can do to put the tags on your site so you should have a real step by step instruction for your big blog networks (and I'm thinking 'click on Layout, click on edit HTML, find the end bit that says , put this bit of code in just before it' or the related wordpress version). I can see the advantage of the ftp being when pages have been set up in php or equivalent and you have to tag more than one page template.

Comment number three (and they are going to get better soon, don't worry). It takes far too long to get to the reports and it times you out far too frequently. I bought my ad center words because I had to. I've still got $5 worth in my account, so that should indicate that I'm not really interested in ppc. Let me log straight into my Analytics account instead of taking me through the whole malarkey of clicking through several pages (ok, its only three). More importantly though, don't keep timing me out so quickly. Google Analytics never times me out. When I open Firefox at work each day, it reloads my last analytics report and I just update it for today.

Lets get onto the positives. Anyone could have copied GA. What Microsoft have done is try to come up with some cool visualisation techniques. Here is one of my favourites - what this shows is each of the pages on my site and their relative traffic (size) and their relative strengths as landing pages (colour).


This looks a bit clumsy for my site - and I happen to have had a load of people come into the home page (we'll come onto that when I look at April's traffic). What I do like though is the way that the content is grouped. You may not be able to tell (and this is something that they could think about doing slightly differently), but each block is actually grouped with other like minded blocks (ie things that follow the structure of my site - all 2008 posts are together, etc).

This 'treemap view' and its related referrer one (see below) are really clever ways of changing how you look at your site. Previously it was a list of pages or a set of content, now we can see the whole and then split down by section. The ability to alter this for pages to show entry pages, pages views, bounces and exit pages as well really helps understand the site traffic. Also my site is relatively small - but for larger sites the zoom functionality looks really cool.


Whilst I am on referrers, the other thing I like about MS Analytics is that when you drill down through your search referrals you don't only get to see which search engine your visits have come from (Google, Yahoo, MSN, etc), but you can see for each whether the user typed in a phrase or came from the home page. This will become more important as the search engines continue to use more images, video, etc (as well as the syndicated homepages - iGoogle, etc). The other nice thing is that as opposed to GA, I can actually see which version of the search engine the users are looking at (Google.com, Google.co.uk, Blogsearch.google.com, etc).
A nice addition to this would be if I could drill from my search keywords the other way to the engines that they came from.

I think to finish off this post I should come up with what is Microsoft's WOW factor (I would call it the USP, but it shouldn't cost you anything) for their Analytics. This is the section that allows you to find out the age, sex, location and occupation of your users based (I presume) on their .Net account. Anyone logged into messenger (this is becoming more common in offices - so don't mock at the teenie generation) or logged into Hotmail who has previously given these details will get this measured. You can't tell who the person is, but I am sure there are privacy concerns. So here we have more of my audience being male than female:


And maybe more importantly for the advertisers, the occupations of all of those users in a way that Hitwise would love:

What do I do because of my Analytics?

The million dollar question is having spent all of last week looking at my analytics, what conclusions have I drawn from it and what am I going to do about it. I think maybe it is a time to revisit last week's analysis and tell you what I'll be doing on my blog because of that. Some things won't tell me anything and some things will lead me up blind alleys, but there is no good not practising what you preach. So here goes:

In March my traffic report showed me that I had most traffic from Organic Search. This seems a good place to start. Now without going into the effort of making sure that my blog is technically optimised, I can still do a number of things to help promote myself on the web. First of all is making sure that all of my posts are linked to from relevent places. That means that I have to ensure that they are all tagged appropriately. They are the labels that you see at the bottom of the post. If I can show to a search engine that I am actually writing in a certain area, then they are going to rank me higher for that area. Hence most of these posts are labelled Web Analytics (hey it's what this blog is about).

How important is this? Well this is the 100,000 dollar question. The answer can be 'not very' if you don't set up your labels into something that a user would search for. Am I good at it? Not so sure at the moment - look at that list down the left hand side and tell me which you might search for. Maybe you'd go for Google Analytics, Persuasion Architecture or Blogging, but I am guessing that you aren't going to search for 'Engagement'. And even if you did, you'd end up finding a load of stuff about engagement rings (and rightly so).

The other thing that this might help you with is people who set up feeds of links on their pages (more web 2.0 news etc). You might not get clicks from the pages, but its good to know that they are linking back in (check in Yahoo site Explorer)

Step number two in this is to ensure that not only do I label everything well, but I use key search terms to link within my posts to other posts. Those of you coming in at the home page won't need to me to link to the next post, but those coming in direct to the post will (they can't see it). This is a good point to pick out some key terms that I want to rank for and see what I can do about it. As you can see, I regularly link 'SiteCatalyst HBX' and 'Omniture SiteCatalyst HBX' back to my original post on the subject of WebSideStory being taken over.

The other article that does well in the search engines is the conversion funnel analysis post. These are key search terms that generate a lot of traffic and appearing higher in rankings will help drive more traffic to the site.

Another good option is to mention company names in your blog posts (and headlines). Many people set up alerts around the web so that they can find out all the blog posts or news around their company. Don't skirt around it (unless you are slamming them, of course).

We've started micro, but we are moving macro now - lets look at which pages got the most traffic at a macro level. I can't split it into labels, but what I can do is split it into age (that is how my urls are structured). If I was using HBX or SiteCatalyts I'd have defined my groups of content so that I could do this by type of content, but for this blog I only have when it was written. It won't surprise you to hear that 2008 got more than 2007. More surprising maybe is that my search labels got more than 2007 as well. Maybe that implies that if I am linking out of the post, instead of linking to old articles, I am better off linking to search headings. This is going to be a trial and error setting, because I don't know the effect it will have. But I can check later.

Drilling down even further I can see that some of my older postings (certainly from January) are now getting very little traffic. Is this because they aren't interesting any more or is it because I stopped writing around them and linking to them. Maybe a revisit to some more of the Hitwise posts are in order. These were popular at the time, but have suffered from the lack of investment in my time.

The other thing I did in March was that I posted more often. Those RSS feeds that I am sure users are looking at (I get traffic from Netvibes and from Google homepages) meanst that if I post more often, I should get more traffic. This means regular posting (at least once a week) and I need to see if I can persuade someone to guest post again (just like Kate did from Search Engine Strategies in New York).

Traffic for Whencanistop

I thought that in this post I might share with you some information on the amount of traffic I've been getting to this blog and give you some insight as to how it works. It should give you a good idea as to what is happening and how I look at my own traffic, having described how to check yours you can see if I am doing it right. More than that, I've just signed up to Gatineau (or Microsoft Ad-Centre Analytics as it now seems to be known) so I'll be able to tell you a bit more about it in the months coming up.

Ok in March I had 253 visits. Averaging 8.43 visits a day, so Google Analytics tells me. Don't believe me? Well here is the graph that says that it is right:


What does that mean? Well lets look at it in a slightly different way, in February I had 86 visits and in January I had 142 visits. So it all looks good for January (January is always great for traffic don't forget) and then goes a bit pair shaped in February. This leads to the question of why this is happening.

Well first step is always to drill down into where people came from. Lets look at January first because it was large. In January I had three big landing pages and using Google Analytics I can drill down into each of these to have a look at where they peaked. Firstly I have the 6 tips for using Hitwise:


As you can see from the graph, one of my big increases in January traffic was caused by this page. Where did they come from? Well in this case I know, because it was an email sent out to our editorial team telling them to read it. You can tell because if you drill down to 'Entrance sources' on the right hand side you can see that.

So we can do that for January, but what about February? I only had one real big landing page and that was fairly frequent throughout the month. In this case it is slightly more interesting because the article was written in January and was about Omniture buying Visual Sciences. So why did it have loads of traffic - this time if you drill down to entry sources we see that this actually got most of its traffic from Google. What we need to do next is to drill down even further in here and look at the keywords and rather unsurprisingly there are a lot of keywords with Omniture and HBX in them. But 51 keywords is a lot for just one page.


So here we can now see that most of the time that I was getting lots of traffic, it was coming from Organic Search. However we should also look at the other referrers and see if there was anything significant in there. Well there were 17 directs and a few from other sites. I should probably go to those other sites and say thank you, maybe leave a comment or link to them from my blog.

Well March is much different. Traffic is much higher. To the naked eye, there was a big increase in traffic around the time that Kate started blogging on SES New York. And actually if you look at the traffic the amount of entries for Kate's take on people being cats and not dogs is around about the same time that I had an increase in traffic. But remember that compare to site functionality because if you do that, it shows that Kate's articles were not the main driver. Lets not get carried away here, those articles have been well read - the cats and dogs one is the fourth most read article on the site in March.

Lets look elsewhere then. The top post was still the article from January about Omniture and HBX. That didn't really peak at any point throughout the month. The home page itself was a bit landing page, but again that didn't really ever peak. The conversion funnel analysis also was in the top three, but that was an old story, so didn't really peak either.

Lets go back to our top 5 tips. Tip one is to look at where people came in to. Tip two is to look at where they came in from. That is much easier - drilling down into the referrers report leaves me to see that whereas in February I had 53 visits from Google, in March I had 149. And those 149 fairly follow the peaks and troughs of my overall site traffic. Good stuff, well optimised content. Lets drill down and see if we can make out why.

Looking at my Google Landing pages shows that I had 30 different pages which people land on. That's surprising because this is only my 30th post. This is where our macro analysis begins. See that url at the top? Its folder structure contains the date that the post was published. Lets drill down into this a bit more - if you look at the top landing pages from Google they are almost all Jan or Feb 2008. Three out of the top ten are March 2008. The top 3 account for about half of the visits, so I am getting a long tail. How do you get a long tail? Writing articles on a fairly regular basis, linking between them in a good way helps. Also making sure that the labeling (or tags) that the post have are well related also helps. No golden bullet on this one I am afraid.

It also helped that in March I got some traffic from the Search Engine Strategies blog (thanks again Kate), not to mention James Kelway's Userpathways blog. Overall a good month and I am hoping that April will be better!!!

Finally, the last thing that you should always do is check out those locations. Most of them came from the US (that's not surprising - that's where most web analytics people are). And particularly actually from New York (thanks for promoting me at SES Kate!!). The second on there is the UK (that's where I'm from, so most of them are probably friends). But actually there are about a fifth of the visits coming from elsewhere. Australia and Germany leading the way at the moment.