Category Archives: Business Marketing Online

Google’s ‘Display Network’ needn’t be overwhelming

Google’s ‘Display Network’ of millions of sites where we can advertise through AdWords has become more attractive as it has offered us more control over the years. As long as you keep it on a tight leash, it can be very cost-effective, even compared to search advertising. There are now so many ways of targeting the ads that using it can be almost overwhelming. Not only can you target pages and sites by name, by keyword and by topic, you can also target viewers by interests, demographics, previous interactions (‘remarketing’), location and devices. And yes, you can combine all or any of those. So where to begin?

We like to start by advertising on known sites, and/or advertising to known people. The former can be done by getting long lists of sites which have referred people in the past – only a few will have Google ads on them, but you can specify them anyway. The latter can be done through remarketing lists. The key is then to monitor things carefully. More often than not we find that mobile devices produce a poorer cost per engaged visitor than desktop PCs, to the extent that we often start without using them, possibly just introducing them as a test later. Other tests we run will include letting the ads run on unspecified sites but with keyword targeting.

There are also different types of ads to test: simple text ads, ‘responsive’ image ads (where Google makes up the ads from supplied images and logos), and your own image ads. While the latter is obviously the most impressive to see in the real world (especially if the ads are HTML5), there’s going to be a limit to the number of sizes you’re prepared to commission, and that limits the reach severely. The key though is to test the results continuously.

Time to have a spring-clean of your page titles

In the eternal quest to find out how to get to the top of the Google search results, we have to consider what the critical factors are. Google won’t tell us, of course – it may not even really know any more – but plenty of people have tried analysing huge data sets to get an idea.

From the research, it’s clear that links to the page and site are key. Typically, the top couple of results for a given search have way more than the rest, so it’s essential to investigate the strength of the pages holding the top positions before you try to target overtaking them. If you’re going to try building links to challenge the leaders, not only do you need to prioritise links from sites which are strong themselves, but the links should either have your search term in their linking (‘anchor’) text, or at least be from pages where that’s the theme.

However, another finding which I’ve seen mentioned a few times is something much easier to implement. It’s to do with having page titles which are concise and on point for the search. Do a few searches for competitive product types (e.g. ‘pneumatic cylinders’, ‘load cells’, or ‘aluminium windows’) and take a look at the leading results yourself. I think you’ll start to see what I mean. It’s not uncommon to see a whole page of results where the titles are just the product, or the search plus two or three words (perhaps the manufacturer’s name). So maybe it’s time to have a spring-clean of your page titles, and to check that they’re really working efficiently for you.

Don’t be put off online display advertising

A single advert placed in a single issue of a good trade magazine can have an impact. It’s a common mistake however to assume that an advert on a website can get a comparable response. Years of results have shown that this is not the case (“What happened? I had an advert running on that website for 12 months and can’t see any clickthroughs”).

Getting equivalent coverage to a print advert from website adverts requires huge exposure. From a magazine, you’d probably be happy to get enquiries equivalent to between a hundredth and a thousandth of the circulation. Online, it’s even harder; if you got enquiries from a thousandth of the advert views, that’d be quite a result. But magazines are offering that response to several dozen advertisers each issue. Few of them have websites which can offer hundreds of thousands of advert views each month. Most can only offer a fraction of that.

To get results from advertising on websites requires far greater numbers of page views than individual specialist websites can ever offer. To be successful, you need to be advertising on large numbers of websites simultaneously. This can only be done through advertising networks like the Google Display Network.

When we run such campaigns for clients, they’re often amazed at how broadly the adverts show. Not only are they surprised to find that the adverts ran on 100 or even 1000 sites, they’re surprised that the sites are quite respectable ones too. If combined with an element of retargeting, online display advertising can be effective and cheap. If you’ve been burned in the past by a media salesperson convincing you that using one website could be good value, don’t let it put you off online display advertising in general.

Google descriptions drop back to two lines (aargh!)

A few short months ago, I wrote about the two-line description snippets in Google results having expanded into four lines. As this change had been slowly appearing for many months, and had now gone full-scale, it seemed like it was finally time to rewrite our description meta tags to take advantage of all those extra words.

Guess what? Google has changed its mind and gone back to two lines, or the old 155-160 character limit.

To say this is irritating is an understatement. Google says: “We were very clear there was no need for people to change meta description tags, when we were asked specifically about this”, but of course many of us tried to take advantage on our most important search and page combinations. When asked what the new character count is, a Google spokesman said: “The count can vary and isn’t helpful in this case to best advise site owners on how to craft meta description tags that might or might not be used for snippets. If we say any character count or any range, it causes obsession over what happens if you’re over or under this supposedly magical figure for a tag that we don’t even always use to craft a snippet. Any specific advice we give will get taken out of context, obsessed over, over-emphasized and optimized for in ways that are not helpful. So I totally get wanting something specific, but we aren’t doing that because it’s less helpful than it seems.”

I’ve held back for a couple of weeks on reporting this change, firstly to see if it’s going to stick, and secondly to see what the recommendations are going forward. It looks like for now we are indeed permanently back to descriptions of a maximum 155-160 characters, so if that’s what you’ve generally worked to with your description meta tags, it’s as you were. If you rewrote some or all of them to double that length, I wouldn’t rush to cut them down – it gives Google more to choose from, and I’ve noticed that if there’s a sentence break around halfway in the expanded description, it often uses that first part on its own, which is fine by me.

Are you measuring views of PDF documents?

Of course it’s great if someone comes on to your website, completes an enquiry form, and you can monitor where they came from. But as I’ve remarked many times, that’s a dubious measure of what the effective sources of visitors to your website are. Firstly, the number of forms completed will never be that high, so the data isn’t going to be that statistically valid. Secondly, many people will come on to your site and enquire in less measurable ways, but it’s not fair to assume the measurable ways are a fair sample. Some sources might be more likely to lead people to make, say, a telephone enquiry than others. Thirdly, many enquirers may be repeat visitors, and the visit where they made the enquiry may not have had the same source as the original visit which was the critical one.

Instead, I prefer to see if there are actions we can measure which:

  1. Indicate that the visitor is interested, and might become an enquirer;
  2. Occur frequently enough to have some statistical validity; and
  3. Are likely to happen on the first visit to the site.

The most obvious of these, and one which we use with most of our AdWords analysis, is simply whether the visitor looked around the site. This passes all three tests, although we accept that it’s a long way from looking around a site to making an enquiry – and whether it can truly represent the latter is of course arguable.

A metric which might be a better representation of genuine visitor interest is tracking the views of sales documents and datasheets. It’s impossible to prove this is a really good measurement of likely enquiries either, but common sense tells us it should be.

So, how do we measure views of sales documents and datasheets? If they’re in PDF or any other non-HTML format, it’s impossible to record them in our analytics application as having been viewed. However, we can measure the click on the web page which opens them. Surprisingly few companies seem to do this; their Google Analytics reports don’t have any record of visitors attempting to look at PDF documents.

This is a real waste, especially as it’s straightforward to set up. Once upon a time, it would require code to be added to each button or link which opened the PDF document. Now, even that’s not necessary. Google Analytics plays very nicely with Google Tag Manager, a system which can make all sorts of things happen behind the scenes in response to visitor actions, including passing their occurrence to another application.

So my questions to you are:

  • How are you measuring a successful visit, and is this the best measure you could be using? and
  • Are you measuring views of PDF documents, and if not, why not?