Measuring web marketing effectiveness

An organisation’s web site offers an unrivalled opportunity to measure significant aspects of marketing effectiveness since customer behaviour can potentially be monitored in real-time and responded to accordingly. Yet research suggests that companies are not making full use of this potential. In a November 1999 report ‘Measuring web success’, Forrester Research asked 50 Global 2,500 companies ‘what metrics do you use to measure your website’s success?’ The responses were:

Hits (82%)
Page views (80%)
Session length (66%)
Visitor count (64%)
Ad banner click-through or referring site (38%)
Look-to-buy ratio (12%)

These figures suggest that even amongst large companies, the potential value of the web site as a marketing research tool has not been tapped. Few companies reported measurement of the site’s role in influencing marketing outcomes such as leads and sales or how it shaped customers’ brand perceptions.

To integrate a measurement programme into web marketing activities is not straightforward. To start with, the demands of promoting a site and keeping it up-to-date may stretch the staff involved with web marketing who often have responsibilities for other channels also. The need to evaluate the web in conjunction with these other channels rather than in isolation also poses technical and organisational problems. As a result it is easy to suffer the symptoms of any poorly conceived marketing measurement programme:

objectives poorly defined;
data not collected;
data not analysed;
results not reported;
corrective action not taken.

Making measurement happen

To avoid these pitfalls, a co-ordinated, structured programme seems necessary. In common with any marketing measurement programme, the starting point is to have clearly defined goals; in this case stating the intended contribution of web marketing to overall marketing and business objectives. Someone also needs to have ownership for the programme – it can be suggested that this role is too important to be entrusted to a more junior member of the team. As Figure 1 shows, once the objectives are established, we can then measure against the objectives (performance measurement); try to understand the variance between what is occurring and our objectives (performance diagnosis) and take corrective action, for example through enhancing the web site or associated marketing communications.

Figure 1 A process for assessing web marketing effectiveness

What should we measure and when?

Once we have a process for our measurement programme, it is helpful to have a framework for different types of measures we need to collect. Research at the University of Derby indicates that, whatever the specific objectives for web marketing within a company, there are generic metrics that are evaluated. In many traditional clicks and mortar businesses, it will be appropriate to base these measures on a traditional approach to performance measurement for a channel such as a phone-based direct sales channel. We need to compare the performance of the Internet channel to other channels and build a picture of how it is contributing to the business.

Performance against targets is mainly monitored in the categories shown in Figure 2. The figure is intended to be read from left to right since achieving the objectives on the left helps achieves objectives to the right.

1. Channel promotion. Promotion is successful if traffic is generated that meets objectives of volume and quality. Quality will be determined by whether visitors are in the target market and have a propensity for the service offered. Overall hits or page views are not enough – inspection of log files for companies show that a high proportion of visitors get no further than the home page! Differences in cost of acquiring customers via different alternative channels can also be assessed. We can use log file analysis to assess which key partner or intermediary sites customers are referred from and even which keywords they type into search engines when trying to locate product information.

2. Channel buyer behaviour. Behaviour objectives will require customers to interact with the appropriate on-site marketing communications such as product information, promotions or customer service. Once customers have been attracted to the site we can monitor, again using log file analysis, content accessed, when they visit and how long they stay for and whether this interaction with content leads to satisfactory marketing outcomes such as new leads or sales. If visitors are incentivised to register on-site it is possible to build-up profiles of behaviour for different segments. Key performance ratios can be identified such as page impressions/visit and visitor sessions/unique visitors.

3. Channel satisfaction. Customer satisfaction with the online experience is vital in achieving the desired channel outcomes, although it is difficult to set specific objectives. Online methods such as online questionnaires, focus groups and interviews can be used to assess customers’ opinions of the web site content and customer service and how it has affected overall perception of brand.

4. Channel outcomes. Traditional marketing objectives such as number of sales, number of leads, conversion rates and targets for customer acquisition and retention should be set and then compared to other channels. Dell Computer (www.dell.com) records on-site sales but also orders generated as a result of site visits, but placed by phone. This is achieved by monitoring calls to a specific phone number unique to the site.

5. Channel profitability. A contribution to business profitability is usually the ultimate aim of web marketing. To assess this, leading companies set an Internet contribution target of achieving a certain proportion of sales via the channel. When Easyjet (www.easyjet.com) launched its e-commerce facility in 1998, it set an Internet contribution target of 30% by 2000. They put the resources and communications plan in place to achieve this and their target was reached in 1999. Assessing contribution is more difficult for a company that cannot sell products online, but the role of the Internet in influencing purchase should be assessed.

The range of measures in Figure 2 currently represents a wish list for most; few companies have yet created the technical and organisational infrastructure. Most companies seem to start by considering volume of traffic (channel buyer behaviour) and outcomes (leads and sales) before moving on to reviewing channel satisfaction and channel promotion.

Figure 2 A framework for measuring web marketing effectiveness

Once a company identifies the SMART measures it wants to collect, the next stage will be to decide on the collection/reporting frequency. For example on-site sales promotions for an e-tailer are often monitored daily or weekly; referring sites may be monitored monthly; while channel satisfaction may only be recorded periodically, perhaps before a major enhancement to the site.

How can we measure?

We can collect much of the information in Figure 2 online. A wealth of marketing research information is available from the web site itself, since every time a visitor clicks on a link this is recorded in a transaction log file that summarises the content the customer is interested in. Since these log files quickly grow to be many thousands of lines long, analysis software is needed to summarise the information contained within them. Log file analysers, of which Webtrends (www.webtrends.com) is the most widely used, highlight interest in products or promotions and how this varies through time (Figure 3). This enables companies to respond in real-time to buyer behaviour. UK e-tailer Jungle.com uses this technique to change the offers on its home page if customers are not responding to a particular offer.

Figure 3 Example WebTrends log file analysis

In addition to information gleaned from the log file, we can also use traditional marketing research techniques in an online setting. Online questionnaires are used increasingly by large companies, for example the Epson UK site (www.epson.co.uk) illustrates how questionnaires can be used to gather information from all stages of the buying process. Interactive tools are available to help users select a particular printer, diagnose and solve faults and technical brochures can be downloaded. Feedback is solicited on how well these services meet customers’ needs. Questionnaires can also be administered anonymously in conjunction with visits by mystery shoppers (Fig 4). Online focus groups also offer potential, but are used less widely than questionnaires. Table 1 presents some of the advantages and disadvantages of these different techniques.

A comparison of different online metrics collection methods

Technique	Strengths	Weaknesses
Log files	· Directly records customer behaviour on site plus where they were referred from · Low cost	· Not based around marketing outcomes such as leads, sales. · Size, even summaries may be over 50 pages long · Doesn’t directly record channel satisfaction · Misleading unless interpreted carefully
Online questionnaires	· Records customer satisfaction · Relatively cheap to create and analyse	· Sample bias – tend to be advocates or disgruntled customer who complete
Online focus groups	· Relatively cheap to create	· Difficult to moderate and co-ordinate

Owing to the weaknesses with the online methods of measuring site effectiveness highlighted by Table 1, online research techniques have to be used in conjunction with traditional offline market research techniques. For example, many companies gather representative users in a room with PCs to conduct a focus group. Mike Bloxham of Netpoll (www.netpoll.net), an organisation specialising in measuring online marketing effectiveness for companies such as Microsoft and Sky.com, believes that conducting real-world focus groups has the benefit that the reactions of site users can be monitored; the scratch of the head and the fist hitting the desk cannot be monitored in the virtual world!

Figure 4 Online questionnaire for completion by mystery shoppers

Real world measurement is also important since the Internet channel does not exist in isolation. It must work in unison with real-world customer service and fulfilment. Chris Russell of eMysteryShopper (www.emysteryshopper.com), a company who has completed online customer service surveys for major UK retailers and travel companies says ‘we also needed to make sure the bricks-and-mortar customer service support was actually supporting what the clicks-and-mortar side was promising. There is no doubt that an e-commerce site has to be a complete customer service fulfilment picture, it can't just be one bit working online that is not supported offline’. An eMysteryShopper survey involves shoppers commenting on site usability, but also on the service quality of e-mail and phone responses together with product fulfilment.

The future

Despite the potential of web site measurement, achieving a unified view of how the Internet channel integrates with marketing communications, sales and service delivered through other channels is still some way off for most companies. Many of the web site analysis tools that are currently used are poor at integrating with other systems for CRM and sales. Suppliers such as Netgenesis (www.netgenesis.com) and Accrue (www.accrue.com) are developing systems that provide a better integrated view of performance. The significant organisational challenges of finding the time to create and manage a measurement programme also need to be met.

Mini glossary

Click-through

A person following a link or graphic from a referring site to reach the current site.

Hits

A hit is recorded for each graphic or block of text requested from a web server. It is not a reliable measure of the number of people viewing a page (or site) since a viewer of a single page may register several hits dependent on the number of graphics downloaded. For example, a web page with 10 graphics would be recorded as 11 hits for a single viewer.

Log files

Log files are present on web servers to record every item requested by users such as graphics and text files. Log file analysers summarise the content of the log file to indicate the usage through time of different parts of a site by different visitors.

Page impressions (views)

A page impression is a more reliable measure than hits that denotes one person viewing one page.

Referring site

The previous site the visitor used before clicking on a link or graphic to reach the current site.

Session length

The length of time a visitor spends on a site (a session ends after inactivity for a set time, e.g. 30 minutes)

Visitors (visitor count)

The number of unique visitors to a web site within a set period.

Visitor sessions

The total number of visits to a web site within a set period.

Biography

Dave delivers the CIM Training seminar Measuring online marketing effectiveness and is an examiner for the new CIM e-marketing professional development award. He is senior lecturer in marketing at the University of Derby and author of Internet Marketing: Strategy, Implementation and Practice.