Long before I came into the technical content marketing and PR business, I was obsessed with marketing metrics. A contributor to that was reading Commonsense Direct Marketing by Drayton Bird and it’s interesting that some of the techniques used to measure direct mail effectiveness can now be applied to PR. As just one example, we can quantify click-throughs from online editorial, and often identify their source.
My obsession stemmed from the frustration of endless discussions (and occasional arguments) about the subjective elements that constituted marketing materials – the main culprits being headlines and colour schemes. It wasn’t a case of what worked best, more personal preference.
Subjective opinions made the process of developing these pieces – whether advertisements, emails, press releases, etc. – long, expensive, and slightly frustrating, especially for the creative team who always felt overly protective of their baby.
Reading the aforementioned book provided me with a moment of clarity: with a bit of ingenuity, almost everything can be measured, alternatives can be tested, and the results evaluated.
All of a sudden, the ability to measure previously subjective issues (such as colour schemes, headline changes, and calls to action) were within reach. But there was a problem. How to determine the source of the traffic, particularly when coming from offline communication pieces – mainly adverts, direct mail and PR.
Adverts and direct mail were easy – we had full control of the content and delivery of these pieces and could use custom URLs on each communications piece. All we now had to do was segment traffic by the URL used to navigate to the website.
But there was an elephant in the room: PR.
We experimented with both custom domain names (e.g. clientnews.com) as well as custom URLs (e.g. client.com/PR-topic), but we couldn’t control what was published. Journalists had the annoying habit of ignoring the links we provided, and using the default website address, or worse yet, not providing one.
So what was the answer? It seems obvious, but we first had to define exactly what we wanted to measure.
What is the purpose of PR?
A simplified view of PR’s role within the marketing function (and specifically tailored towards the AIDA model) is that it helps create / increase awareness, facilitating additional communications activities to drive potential prospects through the other stages (Interest, Desire, Action).
How to measure brand awareness
In order to measure brand awareness, we need to understand the actions that brand-aware visitors are likely to take in order to arrive at your website. It could be one of three:
- Searching for your brand name in Google
- Typing your website address into their browser (or accessing it from their bookmarks)
- By clicking a link to your website from an industry news website (or blog, etc.)
Therefore, to measure brand aware visitors to a website, we need to measure website traffic that is filtered to only include traffic acquired through the above three methods. In order to take this a step further and establish the volume of traffic generated from a particular release/article, this filtered traffic can then be analysed to determine the volume of traffic generated within a specified period of time after a the relevant content has been published.
Measuring visitors from brand searches
Google has made identifying brand searches more difficult by removing keyword analysis from Google Analytics. Keyword searches are available in Google Webmaster Tools, but these approximate figures and are known to be highly inaccurate.
Another route to accessing this information is to segment website traffic by visitors that have arrived through search engines, and have landed on your website homepage. The fact that the visitor has landed on your homepage indicates that they have either search directly for your brand name, or potentially searched for a non brand-related term for which you do not have an optimise landing page. This is likely to be a small proportion of this traffic, but to get an understanding of what this might be, I’d recommend looking at your historical data – before Google removed keyword data from analytics.
Therefore your search filter would be to only include organic visits from search engines, where the landing page was your homepage, for the complete year 2012.
Measuring direct website visitors
Visitors to your website that arrive directly (i.e. aren’t driven through other mediums such as search, social media, referring websites, etc) are a good indication of brand awareness, as they are indicating that yours is the website they want to visit. But there is a potential problem that may skew your figures. Email.
Visitors that arrive by clicking on a link in an email (whether a link in your email signature, or a link in an email marketing campaign) are seen as direct traffic in website analytics programs, unless the links are marked up with campaign tracking code.
The reason for this is beyond the scope of this article (and potentially too boring to discuss within a marketing blog), but as mentioned in my colleague Rob’s recent article on our findings from a recent report created for one of our clients, these can create a large proportion of untracked website visits that are counted as direct visits.
In the case of email marketing campaigns, these are seen as huge spikes of traffic shortly after the email campaign is distributed, but for email signatures, these are likely to be seen as a constant trickle ((unless there is a huge drive from sales / business development teams)) making it very difficult to isolate and remove.
Therefore, in order to accurately identify traffic from emails, and filter these out from our brand visitors, all links need to be tagged appropriately. In the case of Google Analytics, there is a simple tool that can be used to create marked up links: the Google URL builder.
Measuring visitors from referring websites
In recent times, another complication has arisen – online PR. Content is now distributed to, and published by not just the website for various publications, but by blogs and social media too.
So where do these fit into the equation?
We follow the model that PR teams already implement for offline coverage – segment the data by “tier 1” publications – those that are known to have the greatest influence. In order for this to work in the online world, we need to compile a list of the most influential websites and blogs, and segment traffic referred from these sources.
Measuring PR traffic through social media
Social media adds another complication. Many companies publish links to press release through their social media channels alongside additional content marketing activities, such as blog posts, etc. Two options are available here:
- Use the Google URL Builder, as referred to in earlier in this post, and use a URL shortening service (e.g. bit.ly) to reduce the amount number of characters taken up by your tagged URL (especially important for Twitter, where character limits are more restrictive)
- Use a social media distribution service, such as Hootsuite, which will provide individual metrics for each link, although this does have a downside in that metrics now have to be gathered from more than one source.
Creating a correlation between website visitors and PR coverage
So we’ve now been able to filter our website visitors to just those that were brand aware, or referred by a known online source. So how do we establish the correlation?
We need to align and annotate the website data with the PR data – the publication of press releases, articles, opinion pieces, etc.
There’s currently no automated way to do this within Google Analytics – it’s a manual process of adding each entry, so it appears as “tooltip” with the reporting timeline.
At Publitek, we reverse the process. We use an online reporting system that measures all of the activity we undertake, and integrates the coverage received. If we have access to a client’s Google Analytics account, we can then extract the filtered visitor data, and automatically overlay these figures on top of the PR roadmap.