Pay Per Click

Most of my clients advertise in the UK market, but we have recently expanded into the Irish market. My first Irish client was a well known travel agent advertising holidays, so I built a well-structured campaign, put it live and ran it for a week. The first thing I noted was that I was achieving extremely high CTRs; the issue was that I was comparing the CTRs to the UK benchmark (which is about 2.5%). So it made me think: how do the CTRs, both naturally and paid, differ from country to country? Answering this question would give us an excellent insight into the user behavioural differences by country.

I am limited to the amount of data I can get hold of, but here are the results:

What strikes me about these results is that the UK, France and Sweden are all relatively developed and saturated markets for Pay Per Click, whereas Ireland is pretty untouched. The higher CTR could be because of the lack of competition and poor campaigns within sponsored listings, or that Irish users are more likely to click on sponsored listings rather than organic.

The next stage would be to analyse Irish clients who are achieving position 1 rankings organically and take the average CTR compared to the UK.

Philip Pollock
Paid Search Manager

A major part of the reoccurring optimisation process of any good Pay Per Click campaign is the continual addition of negative keywords, assuming that your account includes broad match keywords. This is a never ending task, and is the major downfall of using the broad match type on keywords. Yes it is great to generate traffic, open up your account to new keyword ideas, but you can find that 50-70% of the terms searched for are irrelevant to your campaign, and it can get even worse at ad group level as I’ll demonstrate later on in the blog.

But is Broad Match Modifier (BMM) the saviour of the broad match type?

I am currently conducting an experiment which will revolutionise the way I structure my accounts. Depending on the size of the account and the goals the client wants to achieve, I tend to split down my account very heavily at campaign level, trying to maximise budgets to top performing keywords. This allows me to change the level of investment on keywords which have a very low risk, and a high conversion rate, leading to a nice stable account which converts high. The downside, as there tends to always be one, is that it make optimising, especially the addition of negative keywords, labour intensive and difficult.

The main problem I have is at ad group level. As an example, I am going to take an imaginary client who sells branded clothing. I have broken the campaigns down by brand, then by clothing type. I’ve used all three match types as I want to achieve a broad coverage.

This structure is for demonstration purposes only and I realise it may not be the preferred structure. However, I am trying to achieve a good quality score for all keywords. The problem of the broad match type is that if I typed in PJP Clothes UK, it may not be the clothes ad group which gets triggered. It could be the hats ad group, which would display the incorrect ad and wrong landing page. The way around this is to add negative keywords. But the amount of negative keywords to ensure 100% success is almost infinite.

I think we now need to look at the broad match type in a different view and take it purely to get new keyword ideas and nothing more, so reducing the amount of broad match keywords massively. The new structure would be as follows:

I’ve created a new campaign which only includes the keyword PJP Clothes on broad modifier match type and removed all broad matching from the PJP Clothing campaign. The idea is that there is only one keyword which I need to have to control the related themed search terms. The downside to this is that the CPC will be higher, the quality score will be lower and not all search term will be triggered by this keyword. The positives are a more controlled account, with better controls on budget and you don’t need to worry about incorrect ads being displayed. We are now looking at the term +PJP Clothes purely as a testing area for new keywords, so the benefits of this structure will take time to show.

I’m unsure whether to increase the number of ad groups to the PJP Clothing (b) campaign, but my belief is that it could defeat the whole purpose of having the one related themed keyword.

Philip Pollock
Paid Search Manager