the metrics industry

This page considers the internet metrics industry, looking at the history and commercial basis of measurement businesses such as Jupiter Media Metrix and ACNielsen/NetRatings.

It covers -

• introduction - what is the internet metrics industry?
• the business - how do metrics specialists make their money?
• data sources - where does the information come from?
• trajectory - start-ups, collapses, churn and consolidation in the metrics industry
• snapshots - thumbnails of players such as Nielsen, NetRatings, Jupiter and comScore
• studies

Background information about audience measurement models and challenges is here.

     introduction

There is no agreed definition for the internet metrics industry.

Uncertainty has been exacerbated by the travails of some major players over the past decade. At times they have sought to differentiate themselves from 'old media' (and traditional audience measurement methodologies), claiming that they have capabilities not found among their older competitors. At other times they have sought to wrap themselves in a cloak of respectability and boost plummeting share prices by claiming that they are indistinguishable from traditional audience metrics specialists ... businesses that measured newspaper readership, exposure to highway billboards, demographics of the radio audience and so forth.

Overall we can characterise the internet metrics industry as enterprises that -

• provide information about the 'online population' on a global, regional, national or sectoral basis, with that data being sold to site operators, advertisers and other entities (including academic researchers and government agencies)
• draw on data collected through automated mechanisms (eg toolbars installed by individuals, cookies and even covert reporting tools), online/offline questionnaires, focus groups, traffic data acquired from ISPs and corporate networks
• often provide analysis of what the information means, offering interpretation of historic data and projections about future development (and even offering ancillary services such as SEO).

Some of those enterprises have a global ambit. Others are small and highly specialised. The spectrum of businesses has been criticised for problematical data collection/analysis practice (encouraged by lack of standards), aggressive marketing and a poor record in forecasting. It is thus a common complaint that research from particular vendors is overpriced (with substantial cloning of previous data), is opaque and cannot be reconciled with comparable data from competitors, leading one corporate consumer to characterise much commercial metrics as "faith based but not science".

Despite claims of uniqueness the industry has increasingly come to form part of the broader market analysis and audience tracking sector, with colonisation by businesses and personnel with a background outside the web.

That colonisation reflects the institutional contacts and expertise of 'old media' metrics providers, along with the perceived low cost of acquiring dot-com specialists after the 2000 Crash. It also reflects recognition that there is little functional difference between identifying offline audiences and online audiences, ie the essential task is identification and analysis (irrespective of medium).

     the business

The web metrics (or internet metrics) industry centres on sale of data about the size and shape of online audiences.

As noted above, some data is historic: a picture of the audience last week, last month, last year, last decade. Some is predictive: forecasts, based on that historic information, about the nature of future audiences.

Information may be very granular (eg statistics about the number of people in a particular demographic visiting a specific web site) or quite abstract (eg the number of people 'online' in a particular country or the gender ratio of the 'wired population' in a region such as Europe or South America).

Information is often comparative, providing a national or sectoral ranking of sites (eg a 'top one hundred' list of the 'most visited' sites or of the sites 'most visited' by a particular demographic) or benchmarking of a particular site's performance in relation to peers (x competitor had y% more visitors who remained on that site z% longer).

It may be contextual, eg that x% of visits came from a particular region or had previously visited z% category of sites. It may be quite rich, featuring information about age, income, education, ethnicity or other attributes of visitors.

The industry exists because there is a market for a range of data. That market is very similar to the market for offline audience research data. It encompasses -

• advertisers seeking a sense of who is online (or is likely to be online in future), for example to determine whether investment in marketing is justified
• government agencies, academic institutions and other entities wanting a picture of online populations
• site operators (businesses, nonprofits, government, educational) wanting information about who is visiting their sites - and the nature of the interaction - and visiting comparable sites
• advertisers interested in identifying the overall/sectoral popularity of sites, as the basis for paying site operators to feature advertisements (a model used in broadcasting, where networks charge higher rates for ads in the most popular programs and where advertisers often choose to spend their marketing budget at the stations with the largest audience).

Responses to that demand vary considerably.

Some metrics services have offered detailed reports on the size and shape of an overall online population. Others have specialised in particular demographics, with several specialists for example concentrating on ethnic, age or affinity niches. Some have packed reports covering all major sites, focussing their marketing efforts on sale of raw data and analysis to advertisers and other end users. Other services have emphasised collection of data on behalf of site operators, providing information for organisations that do not want to undertake collection inhouse and have an interest in data about comparable sites that is collected by the same service.

Some output from metrics services is consistently quantitative and detailed; other reports may take the form of abstracts and forecasts that have a qualitative basis (often derived on a 'black box' basis using proprietary methodologies and undisclosed raw data). One result is the discrepancies highlighted in the preceding page of this guide, where competing services claim to provide an authoritative account of the same population but offer quite different information - information that cannot be independently verified by a third party and may even be inconsistent with a site operator's inhouse monitoring.

     data sources

The data used by metrics services typically comes from one or more sources, several of which are indicated below. That data may be analysed using different methodologies and integrated with information about 'offline' attributes, some of which is provided by information brokers.

Sources include -

• aggregate information sold/given to a metrics specialist by an ISP (eg identifying how long its average subscriber was online, which sites were most visited, the duration of visitation and even sequences of visits)
• aggregate information provided by site owners about visits to those sites, often on the basis that provision entitles the operator to receive a discount when purchasing data about peer sites and/or the online population
• email, phone and web-based surveys by metrics services about the respondent's age, income, education, gender, experience, consumption patters and perceptions of a specific site or class of sites
• observation by metrics specialists of how users navigate particular sites, sample population focus groups seeking quantitative and qualitative information about specific sites or broader responses (eg perceptions of whether internet shopping is safe)
• toolbars (such as Amazon's Alexa) installed by internet users and providing the toolbar promoter with information about what the user visits
• covert audience tracking mechanisms, eg spyware that was not deliberately installed by a user and of which the user might never be aware.

     studies

There has been no comprehensive independent study of the global internet metrics industry or activity in Australia. Pointers to relevant resources are provided in the discussion of audience measurement elsewhere on this site.

   next page (visualisation)