Google AdWords have released a new report for monitoring the search queries that trigger clicks on your adverts (I found this via SEOmoz).

I have just been playing around with this report on a couple of our client accounts and it’s brilliant.

Much of the data could be found by trawling through your server logs to find the actual queries that are leading people to click on your adverts when you do phrase or broad match, but I know I’m guilty of not spending enough time analysing logs in these days of powerful analytics software.

I have already found a few negative keywords that I should be excluding (I also found out that PSP) is an airport code as well as a handheld console and a piece of software - not sure that most people searching for cars and PSP are looking for car rental from Palm Springs - I reckon more are looking for computer games…).

Now, this new report shouldn’t really lead me to find these new negative keywords, but as I mentioned - with Google Analytics being so good, I’m not concentrating on log files as much as I should be. So this got me to thinking - why can’t Google Analytics offer me some variant of this report?

The ‘traffic sources’ report in GA allows you to filter by ‘paid’ or ‘non-paid’ search and then subsequently by keyword, but if you filter by paid search, the keywords it displays are the keywords you bid on rather than the actual searches people carried out to find (and click on!) your advert. Why is this the only option?

Please include real search queries in the paid search monitoring

Dear analytics development team,

If you’re listening, please can we have a report within analytics that lets us compare the search phrases we bid on with the actual queries people carry out to find and click on our adverts?

Thank you,

A hopeful search marketer

PS I have found that there is a convoluted way of doing this, but it’s pretty advanced, and I see no reason why GA and AdWords couldn’t play nicely together here by default? Especially now you are making this information available in an AdWords report

PPS incidentally, why does the AdWords report hide data on some low volume queries? Why not let me decide for myself what’s significant and what’s not? Given that I could find this information myself from my logs, you’re not keeping secret any information that isn’t out there anyway, so where’s the harm in making it easier to find?

Tweet This Post 

DaveN points to an upcoming release of an AdCenter accreditation system similar to getting ‘Google Qualified’ status.

I like a lot of what AdCenter is doing at the moment and the name of the program (adexcellence) gives me hope that it will be more than a ‘do you know the rules and the basics’ type of test, and more of a distinguishing qualification that really means something.

The tools and research that Microsoft are making available are great and we are getting some really good ROI in the US running AdCenter ads.

My only complaint about AdCenter

The online interface to AdCenter contains some of the most annoying uses of ajax I have ever seen. When I first started using it, it didn’t work at all in Firefox, so I had to fire up Internet Explorer to get anything done (quite annoying when you’re working on an Ubuntu machine.

The other week, I was going through my usual palaver of using Wine to run IE so I could edit some AdCenter ads and discovered that the ajax wasn’t working so I couldn’t actually get at my currently running ads. At first, I thought that it must be because I was running a weird Ubuntu version of internet explorer, but I got the same problem in IE6 on Windows XP. I tried logging in with Firefox and everything worked perfectly.

Please Microsoft, sort out the cross-browser stuff so that AdCenter can become my favourite ad platform…

Tweet This Post 

Something a bit light-hearted for a Friday evening. A little while back, Rand at SEOmoz wrote about the headline-writing lessons to be learnt from reading trashy magazines (I think, though I can’t seem to find the post now).

We’re talking here about such distinguished titles as:

• heat
• zoo

We were talking here at Distilled about how (while we might arguably be writing about interesting things) our headlines left something to be desired and so I have been looking out for great examples of headline writing. This is not a new skill - newspapers and magazines have been honing this skill for over a century (and our great writers have been at it for even longer; I’m sure that Shakespeare would have preferred to have his skills advertised in a magazine “Juliet’s secret man is a capulet” than a worthy review). And there is no doubt about it; it is an incredibly skillful task.

An acquaintance of mine works on the newsdesk of a national newspaper and, over a few beers, as well as telling stories of famous people up to naughty things (that they know are true, but don’t have quite enough proof to print), he enjoys regaling us with the quick wit of the subs who write most of the headlines. These guys are lewd, competitive, hilarious and lightning-quick. Some of the headlines they can’t use are absolutely brilliant.

Online, we are much less constrained about what we can and can’t write, but unfortunately, we also aren’t aided by a team of witty sub-editors to create brilliant headlines for our writing.

I am trying to improve my headline-writing skills by paying attention when something grabs my attention: in print, on an advertising hoarding, on someone’s website or blog, or on digg. What exactly is it that made me sit up and pay attention? (Incidentally, I’m also trying to hone my eye for design by actually paying attention to what it is about some designs that works - next time you see a pretty product or advert, try to break it down into its constituent parts and you’ll see a lot of the same elements repeated - much like the Web 2.0 design rules - if you haven’t read this beauty from web design from scratch, go and read it now…).

A reputation monitoring challenge

To tie this all back to our core topic, of reputation monitoring, when you are trying to keep up with all the chatter about your company or brand online, you are generally going to be presented with a list of page titles that may or may not be about you. Here the challenge is a slightly different one - more akin to speed-reading than looking out for eye-grabbing headlines. You want to read the stories that are about you regardless of headline, rather than reading the ones that catch your eye because of some clever titling. We have been trying to improve this process in our product, by including snippets from the source to enable easy scanning, but there is more to do.

One feature request that a few people have made recently is for ‘importance’ weighting in order to get an immediate idea of how important the blog or website is that mentions you. We have brainstormed a few metrics for doing this, and it may well make an appearance in some form in future releases, but the fundamental problem is that it slightly goes against the core belief we have about doing online blog monitoring in particular - that even if it isn’t an ‘a-list’ blogger who is writing about you, you do need to know about it (especially if the coverage is negative) because it can balloon out of control very quickly.

If techmeme mentions you, you’re pretty likely to know about it, but if an obscure blogger raises a genuine problem, then finding out about it and having the opportunity to respond is hugely valuable - and I’m worried that we might lose these in the noise if we assign ‘importance’ ranks to the sources… Probably something to explore in more detail another day.

Tweet This Post 

I just read Kelvin’s post on Marketing Pilgrim about Gordon Brown’s reputation management. Worth a read.

We have toyed with doing a reputation wars between two politicians, but we need to improve our tools a little bit first. At the moment, reputation monitor is slightly better suited to monitoring companies and brands than high profile individuals (unless those individuals live online and have an obvious website people link to when talking about them - it works pretty well for tracking mentions of me, for example).

The main problem we face when tracking mentions of companies and brands is whether they are actually about the company or brand we want to track rather than something else with a similar name. When you are talking about someone like Gordon Brown or Tony Blair, most mentions of someone with that name are going to be them and the challenge instead becomes working out which mentions are particularly significant and also how positive or negative they are.

Sentiment analysis revisited

Automatically understanding the sentiment of a mention is something I have talked about before. It’s something we would definitely like to get into reputation monitor. We are working on it and we’ll probably write more here about the challenges as we go along. Wish us luck!

US Presidential election reputation management

Also on Marketing Pilgrim, is a reputation study they have done on the 2008 Presidential Election Candidates which looks interesting. I don’t know how they pulled that together, but I’m betting a bit of automated sentiment analysis would have helped!

Tweet This Post 

I met Louise Third from Integra Communications at an event recently where she was speaking on the subject of DIY PR. I was very impressed with her energy and enthusiasm for small businesses and getting coverage for all. We got talking and she mentioned an ebook that she had recently written. It’s called PR on a beermat - part of the beermat series.

I bought a copy a couple of days ago and have just finished it. It is full of good advice, covering subjects from “why should I do PR”, through “can I do my own PR?” and deep into some of the details of how you can go about running a successful campaign yourself. I have found Louise herself very inspiring - she has already given me the confidence to speak to journalists on national newspapers and leading industry journals (more of which soon).

I particularly liked the specific bits of advice such as ideas for creating news, what time to speak to various journalists and the best ways of following up press releases. All of this will be second nature to seasoned PR professionals, but for someone like me on the outskirts of the industry, it is hugely valuable.

Online PR

It would be wrong to write an ebook on a topic like PR without going into the online elements, and PR on a beermat has a good section on the ways the online world is shaping PR. This is an area that is of particular interest to me, given our upcoming seminar on online PR and the campaigns we are running aimed at UK PR firms and Louise handles it well. Her advice on blogging is sound and includes a good discussion of the difficulty of growing an online presence, but also the benefits if you can achieve it.

Her blogging case study is wiggly wigglers for whom blogging has apparently been a great business boost. It isn’t a blog I had heard of (I don’t think I’m quite in their target market) but if you want to read about lacewings, it looks like a good place to start.

Web 2.0

Web 2.0 even merits a mention and I think the following quote from the ebook sums up digg brilliantly:

Often attempts at web PR turn into self-indulgent exercises in vanity - personally enjoyable perhaps, but potentially damaging to the business

Media monitoring

There is a brief mention towards the end of the ebook about cuttings services and how they are almost certainly too expensive for most small companies doing their own PR. I’m hoping that the next edition might have a mention of Reputation Monitor as a tool for doing online media monitoring which is affordable even if you are a small company doing it yourself!

[Disclosure: I have met Louise Third, the author of PR on a beermat, and we got on well. I bought PR on a beermat via their website at full price, however, and have nothing to gain from further sales, so go buy a copy]

Tweet This Post 

Ok, so I’ve been busy and not really had a chance to post (also - I’ve been uninspired). I’m keeping this post short and sweet, so let’s start by talking about Raymond Chandler. Say what? Well - I love reading Raymond Chandler books and sometimes I like to think of myself as the online equivalent of Philip Marlowe (the lead character from his books)

Combining suave sophistication, witty put downs and online investigative powers which border on the super-human. Ok, so I don’t have much sophistication and I’m not that witty (plus I don’t get to sleep with as many of my clients as he does. Disclaimer: ANY of my clients!) but I AM pretty good at the online investigation. You know the sort, where a new client comes on board, they ask your advice on an issue, you come up with a solution in a few minutes by using the online tools at your fingertips, some complex Google searches and a whole lot of know-how.

Now often, this is just a theoretical solution and implementing it can be much trickier but doesn’t it make you feel good when you solve these problems? Just me? Ok, well I love the challenge of a fresh problem, even more so if I can solve it quickly and help someone out which is why:

I’m answering your questions for free

That’s right - just email me at tom@distilled.co.uk (the same address for all fan mail ) and I’ll take a look at your problem/question. If I can’t answer it, you can have your money back!

Alternatively, just leave a question in the comments.

Please - I’m looking for something which has been troubling you, a nagging issue, burning question or flummoxing results.

Please don’t ask me: “How do I rank 1st for this phrase” - I’m not giving away the whole playbook here guys.

Please don’t ask me: “What’s an XML sitemap” - you can all find wikipedia (hiss) yourselves.

This offer is open to webmasters, bloggers, young, old. Heck - I’m even offering it to any rival SEO firms who internally are scratching their heads about a problem (confidentiality will be preserved, just mark your email confidential and I’ll treat it as such).

Disclaimer: I’m not the most incredible SEO’er to walk the earth, I’m no Rand, Grey Wolf, or Danny Sullivan but hey - they all charge for their services so go figure

PS - It appears Raymond Chandler may have predicted Google many years ago, read about it here!

PPS - After writing this article I have seen this post from Grey Wolf. I realise this is a similar idea but really it’s not since I’ll answer specifics for you (he’s looking for more general discussion I believe). This post wasn’t based on that at all!

Tweet This Post 

On Wednesday 11th July, we are running a seminar at our offices at Waterloo called How to make PR work online. The seminar will last for 2 hours, with a break for some nibbles, a glass of wine or a beer and a chance to network.

It’s targeted mainly at agencies and people working in PR, though it’ll hopefully also be interesting to people doing their own PR as well. We are not intending to talk about how to do PR or the fundamentals of good PR campaigns (the people this is aimed at know far more about that kind of thing than we do) but rather the plan is to share some of our knowledge about translating PR into online benefit (and particularly into benefit with the search engines).

We have gained insight into best practices through working with PR agencies, watching PR go wrong (and go right) online and through understanding what the search engines are looking for when they judge the popularity and visibility of your company or brand (or your clients’ companies and brands).

The seminar costs £65 +VAT per head and you can sign up online.

Finally, as regular readers will know, we recently released a tool to help companies monitor anything said about them online (Reputation Monitor) and we are offering a free seminar place to anyone signing up for an agency account before the day.

Tweet This Post 

As I was walking through the station yesterday, my eye was caught by the headlines on The Evening Standard, the paid-for newspaper that competes with the various freesheets in London (thelondonpaper and whatever the other one is called). It often annoys me that their headlines give you just too little information. I know that they’re not running a public service, but I could do with a newsstand that just told me a few headlines on my way to the train so that I wasn’t completely out of touch. I’m not really looking to read a newspaper at the end of a day at work - if I don’t have to read something work-related and Duncan and I aren’t having an impromptu meeting on the train home - I tend to read a novel to relax.

They obviously do it deliberately to try to encourage people to buy the paper, and I’m sure they have tested different approaches and have concluded that it is better to leave crucial information out than trying to entice people in to reading the rest of the story by giving away a bit more of what it’s about.

Using age-old tricks online

In the online world, a lot of the headline writing we do (especially while blogging) is aimed at getting a click rather than a sale, so the barrier is even lower than digging a couple of 20ps out of your pocket as you walk past the Standard newsstand. This should mean that this approach works even better online. Next time you are doing some headline split-testing, why not try failing to tell the whole story?

Tweet This Post 

I’ve been doing some quite fun stuff involving advanced maths over the last couple of days. It’s good to do hard theoretical stuff from time to time. Because it’s been a while, I had to revise my stats quite a bit. As soon as I got beyond binomial and normal distributions to chi-squared distributions, G-tests and beta functions, I turned to the textbooks. Actually, first I turned to Google, but all I could find was course listings for university courses saying things like ‘This course includes the beta function, chi-squared distribution etc. etc.’. I guess most of the notes are only available on intranets etc. It is remarkably hard to find information online about advanced maths.

Wikipedia is actually a very good resource here (this is the kind of thing wikipedia generally rocks at - the right answer is indisputable, so you don’t get edit wars etc.). To see how comprehensive it is, check out the pages on the Beta function and its distribution (though there’s precious little about how to calculate it).

When it comes to numerical methods, there’s still nothing to beat Numerical Recipes but I didn’t really want to have to write my own routines to calculate things like chi-square distributions. Luckily, after a bit of googling, I found that we could install statistics functions for php. Unfortunately, they are the worst-documented functions I have ever seen (check out the documentation for the chi-square distribution function!) but that’s another blog post.

Why I am looking up all these ridiculous things

I’m in the process of writing a new tool to help us do better at ppc management - a similar idea to Clickmuse’s Adwords Optimizer, that we use and find handy - but instead of telling you when one advert is out-performing another, this is designed to spot keywords that need to be targeted differently (either moved into their own ad group in order to gain the benefit of a high-performing keyword or deleted / moved out of an ad group to avoid pulling the performance of your other keywords down). With the way that quality score is calculated, taking into account ad group performance as a whole and factoring it into your cost-per-click via the quality score, it is becoming ever more important to monitor ad groups closely and group keywords together even better.

At first, I thought this was a pretty simple statistical task - all I needed to do was look at each keyword and its clicks in order to calculate the probability of getting this many clicks if the keyword has an underlying long-run click-through rate equal to the ad group it is in. If this probability was low (below 5%, say), we could be relatively confident that this keyword should be moved out of this ad group as it is an outlier compared to the rest of the group.

I started coding this without thinking too much further (great software development process there) and hit a hurdle when I needed to calculate the probability of getting at least n clicks out of I impressions with a supposed click-through rate of p. It’s simple when I is small (it’s just the sum of binomial distribution probabilities) but as I grows, this involves calculating factorial of a bunch of large numbers. At this point I started dredging up some of my old stats courses and remembered that you can use Normal and Poisson approximations to the Binomial distribution for large populations (which you use depends on the probability - for reasonable size expected numbers of clicks, you can use the Normal distribution, for small expected numbers of clicks you have to use the Poisson distribution).

Calculating the normal distribution in PHP was pretty straightforward, but calculating Poisson distributions when large numbers are involved means calculating Beta functions and other such fun. I needed to read up on the subject. Hence why I turned to Google, failed, and turned to wikipedia.

This post isn’t really about what the solution to my problem was in the end, but just in case anyone is interested, I have gone with calculating a G-statistic for the set of data of:

This statistic then has a chi-squared distribution with 1 degree of freedom (I think - because there are 2 independent variables given the total number of impressions and we are estimating one parameter - the click-through rate of the whole ad group). Testing this for significance then tells us the probability that the keyword clicks and ‘rest of ad group’ clicks are drawn from a universe with the same click-through rate. I think this is a better approach because it allows for both the possibility that the keyword we are interested in has an unexpected number of clicks and that the rest of the ad group has an unexpected number of clicks.

Where this all fails

I find the concept of degrees of freedom very difficult to grasp (always have, despite having it explained by some of the best minds in the world - I find it especially hard when you start talking fractional degrees of freedom) and when I’m working with real-world examples, I’m always sure I’ve got it wrong.

Wikipedia is very poor on the subject and I can’t find good explanations online.

If you happen to understand this kind of stuff and can tell me whether my problem has 1 or 2 degrees of freedom, it would be very much appreciated!

Tweet This Post 

I had a post all written asking “Does all the stuff that has been written about webmaster central apply to all the countries they’ve rolled it out to?”.

We had been seeing, across our own website and a number of our clients’ websites in the UK, a woeful under-counting of links. In the interview with Matt Cutts that Rand did at SMX, there was a discussion around providing good link data and both Rand and Matt seemed happy that webmaster central was giving that now. At the time, I logged in, had a look round and left (again) as it was not showing interesting data for our UK-based sites.

For example, one of our clients has a page with over 100 inlinks according to Yahoo! (we saw most of them arrive, through Reputation Monitor since it’s a new page so we know they are real) but had none according to Google’s webmaster tools. Now, we know that Google knew about those links really because some are highly topical, timely and relevant and as a result, Google is ranking this particular page quite well.

So my post was a big “what’s going on, Google?”, but as I just logged in to find some numbers to back up my argument, I discovered loads of data! All of a sudden, a bunch of our sites have useful data across all pages. Suddenly, I realise what all the fuss is about and the distribution of links, in particular, to different pages is a very nice way of seeing how all your pages are faring.

I think I need to polish my tinfoil hat. The borg is watching me…

Tweet This Post