NRS: Joining up the dots between on and offline audience measurement
Katherine Page, technical consultant at the National Readership Survey, explains why it is worth pursuing Project Fusion. “All the indications are that publisher websites have considerable potential to extend audience reach…”
It’s becoming increasingly important to join up the dots between the measurement of on and offline audiences. For publishers there’s also a growing demand to know how print and online fit together, and what other emerging platforms, such as mobile, will add to reach. This has become a key development objective for the National Readership Survey (NRS).
Publishers need to know how their overall on/offline audiences are developing, and the relationship between the two. Furthermore, there is considerable scope to develop brand display advertising online – this type of advertising still accounts for just 5% of total online expenditure, according to the IAB.
As ever, coherent and trusted audience estimates will be an important tool to develop the market. At the same time, there is a frequent cry that there are already “too many numbers” in the market, particularly as those numbers don’t always tell the same story.
There are two main kinds of measure already in the market: site-centric and user-centric. Site-centric measures are the website analytics, such as Google and Omniture, which provide a census count of activity for particular websites. It is these measures that ABCe audits, when commissioned to do so.
These measures are counts of devices rather than people. It isn’t a straight equation between people and devices – in some cases more than one person will access a site from the same computer, and in other cases the same person will access the same site but from two different devices.
The other thing to remember about site-centric analytics is that they are based on cookies. Some Internet users delete their cookies and the computer concerned will be counted as a new device each time it goes back to the site with a fresh cookie. ABCe has switched the focus from a monthly to a daily count in their audited reports to avoid the risk of artificial inflation from cookie deletion.
The other type of measure is user-centric, based on panels of respondents who have their software installed to track their computer and Internet usage. This is the type of measure provided by UKOM/Nielsen and comScore respectively. These panels do measure people rather than devices, and can provide audience profile data on a like-for-like basis. UKOM formally launched earlier this year to provide an industry-agreed planning system, building on and developing the existing Nielsen panel.
But the challenge for user-centric measures is sample. Large panels are required to measure the long tail of sites, and realistically these cannot be of the same quality of the probability sample used by the NRS. Furthermore, it is particularly challenging to recruit samples to measure Internet usage at work.
It is for this reason comScore is now using a “hybrid” methodology for some of the sites it measures, combining user-centric panel data with site-centric data. In this way, comScore aims to retain the strengths of user-centric measurement, while compensating for the known weaknesses of the traditional panel approach in representing out-of-home usage. This method has its own challenges and provides yet another set of figures, of course.
For the NRS, the issue is how to provide a measure of website audiences in conjunction with the ongoing print readership estimates. This will provide audience reach and profile data for publications and their websites, separately or in combination, so as to assess the net reach of any desired combination. The end objective is to provide a single integrated database for planning across print and digital platforms.
The NRS has already carried out a number of experiments in this field, but there is a limit as to what can be achieved by asking respondents to recall the websites that they have visited. Respondents remember sites that they visit frequently, but won’t remember all the sites they’ve visited, and all of the visits that they have made, particularly when they come to a site in an indirect manner.
The NRS is also very conscious that adding yet another set of figures to the market is less than desirable, even setting aside the costs of generating yet another set of online audience estimates. Hence the creation of “Project Fusion”, which aims to make best use of existing online audience data, while ensuring that the quality of the print readership estimates is ring-fenced.
In essence, Project Fusion will involve fusing website audience data from either UKOM/Nielsen or comScore onto the NRS database.
Of course, there are considerable challenges for such an exercise, not least reproducing an appropriate level of duplication between the print and online audiences for each publisher. This will be the key to the success of the project if the combined database is to be used for reach and frequency planning.
With that in mind, the first stage of the project will be an experimental fusion later on this year, which will enable NRS to assess the quality of the fusions.
Despite the undoubted challenges, it is well worth pursing Project Fusion – all the indications are that publisher websites have considerable potential to extend audience reach.