ellipsis flag icon-blogicon-check icon-comments icon-email icon-error icon-facebook icon-follow-comment icon-googleicon-hamburger icon-imedia-blog icon-imediaicon-instagramicon-left-arrow icon-linked-in icon-linked icon-linkedin icon-multi-page-view icon-person icon-print icon-right-arrow icon-save icon-searchicon-share-arrow icon-single-page-view icon-tag icon-twitter icon-unfollow icon-upload icon-valid icon-video-play icon-views icon-website icon-youtubelogo-imedia-white logo-imedia logo-mediaWhite review-star thumbs_down thumbs_up

The Best Website Metrics Are...

The Best Website Metrics Are... Brandt Dainow

Web analysis is often seen as a group of disjointed tasks, each looking at a different aspect of a site's activity. Marketing performance is analyzed in terms of impressions and clickthroughs, while sales performance is examined with metrics such as shopping cart abandonment rates and average sales values.

In my opinion, this approach obscures rather than helps. It is used because that is the way the data is gathered, and few developers have paid any attention to presenting the gathered data in a way that can facilitate an understanding of the bottom line. Yet it is the bottom line that determines whether a site is successful or not. In addition, these traditional forms of web analytics data presentation do not give me the clear, easy-to-understand picture I need for reporting to senior management.

I propose a different approach, one in which each stage is seen as merely one step in a continuous process. This process commences with an ad impression and concludes with a credit card or contact form submission. I call this approach Integrated Path Analysis.

Many people already see web analysis this way, and it is implicit in much of what is taught. This article simply seeks to make Integrated Path Analysis explicit and obvious.

Author notes: Brandt Dainow is an independent web analytics consultant and the CEO of ThinkMetrics. Read full bio.

The prospect ratio is rarely examined, yet it is of critical importance. At a glance the prospect ratio tells us how successful the site is as a sales tool. It has an even more critical role once we have acquired the conversion ratio, because this can tell us if our final form design is OK.

The conversion ratio is officially defined as the percentage of total visitors who complete the target action. In practice I have come to regard this definition as too vague for practical application, since it includes people who bounced from the landing pages.

When I am analyzing what happened inside a site, I don't want to include people who didn't make it beyond the landing page. Integrated Path Analysis looks at the percentage of prospects who went on to completion.

To avoid confusing this with the traditional standard for conversion rate, we call this the "prospect conversion rate." This is the reciprocal of the officially-defined abandonment rate. The abandonment rate is the percentage of people who put things in a shopping cart but bail out at the credit card page, or who view a contact form but do not fill it in.

Shopping cart abandonment rates are typically between 40 percent and 50 percent, while contact forms can be up over 80 percent or as low as 20 percent. Our prospect conversion rate is the percentage of people who did not abandon.

Multiplying the cost-per-prospect by the prospect conversion rate provides the most important number possible in any web analysis: the cost-per-acquisition (or cost-per-sale). We absolutely need to know cost-per-acquisition for any website. If it is too high, it will eat up all our profits and we'll go out of business. The lower it is, the better.

Deriving a conversion ratio based on those who saw the target action page tells us immediately about the quality of the design of that page. The people we are measuring here are those who had qualified in terms of meeting all the criteria but didn't commit.

While not every one of them abandoned the page because of the design of that page, improving that design will always increase the percentage of conversions, often by a very large amount. It is not uncommon to see improvements of 300 percent or more from a single design change. Since this is the final measure of sales, improvements in the conversion rate translate directly into increased sales.

By working at the end of the path, the conversion rate becomes a fulcrum by which we can lever the overall sales success of the entire online business.

Analyzing what is going on inside the site can seem overwhelming. It is the place where most web analytics products will give you the most information. Unfortunately, for most purposes that information is useless.

We're not interested in how many page views it took people to get to this point, how long it took or what the most popular pages were. Such information is so broad, so top level and contains so many contradictory elements that it is actually misleading. It is only of practical use when applied to a single page or section of a site, and we're trying to determine how people are interacting with it. So my advice would be to discard or ignore much of the data available at this point. 

Visitors will do many things inside a site. Almost every visitor's path is unique. People will look at some pages for minutes and others for mere seconds; they will reverse back through a series of pages and revisit something, then they'll move forward, back up and generally confound any attempt you could make to find common patterns of behavior.

It doesn't matter because we never need to look at a site that way.

For the purposes of Integrated Path Analysis, all that matters is what percentage of readers end up looking at a contact form or at a bunch of items in a shopping cart. What happened between the landing page and that point is irrelevant. There is no standard terminology to define people at this point, but we need to measure them, so we can call such people "prospects."

Integrated Path Analysis doesn't care what, why or how people changed from readers into prospects, simply what the ratio is between the two.

The percentage of readers who become prospects is our prospect ratio. By combining our prospect ratio with our cost-per-reader, we can determine the cost-per-prospect. Using the earlier example, if 25 percent of the people who read the site put items in their shopping basket, we have a cost-per-prospect of $15.40.

At this point (and not before) we have a potential sale.

A website is pointless if no one goes to it. In the discussion that follows, we will focus on visitors arriving via online advertising, but Integrated Path Analysis works just as well for native search listings, affiliate programs or simple link exchanges.

Ad presentation
The most important number we need to understand the start of the visitor's path is cost-per-visitor (CPV), or cost-per-click (CPC). If you are buying pay-per-click ads, this is presented to you in your account reports. If you are paying for impressions, then you need to calculate this from the total spend and the number of visitors generated.

For the purposes of Integrated Path Analysis, we are not concerned with the number of ad impressions or the clickthrough rate (CTR). Those factors are important for assessing and improving the performance of the ads, but that is not the point of Integrated Path Analysis.

We are concerned with the critical decisions that move someone toward purchase. The first critical decision a potential customer makes is whether to click on our ad (or not). For our purposes, we're not very interested in the number of people who clicked, but simply in how much we paid for each click.

Landing pages
Visitors move from ad to landing page. If you have not provided different entry pages specifically designed for different ads, you will probably be using the homepage as your generic landing page.

Almost all visitors decide whether to leave or enter your site based on their assessment of the landing page. They use this page to discover the content of the site, and to assess the navigation system and other functional elements to see how it works. In this they are trying to decide if what they are looking for is contained within the site, and whether they will be able to find it without too much effort.

The critical thing about landing pages is that it costs money to show them to people. This is the cost-per-click we determined in the ad presentation section above. If people get to a landing page and then bounce (or leave), it still costs money. This "bounce cost" has to be recovered later from sales.

In other words, our customers subsidize our bouncers. This subsidy is factored into our analysis at this point by determination of the cost-per-reader. A reader is a visitor who enters the site, rather than bouncing. Cost-per-reader is determined by combining the CPC with the clickthrough rate for our landing pages, thus:

In the above example, we are paying $2.50 per click in our advertising, and 35 percent of people who click the ad leave without going beyond the landing page. The net result is that it costs $3.85 to get someone to read the site.

In other words, this is what we are paying to commence our sales pitch.

What we are seeking to do is to compress our mental vision of the path taken from initial ad exposure to purchase completion. The clearest indications of website performance are achieved by working with these two ends of the process. By creating a compressed end-to-end view, Integrated Path Analysis aims to make this connection obvious and thus easier to work with.

Integrated Path Analysis only concerns itself with the metrics that reflect critical points in the visitor's journey through our online material. These critical points are ones in which the visitor must make a yes or no decision as to whether to proceed further down the path or terminate the path (by leaving). Integrated Path Analysis is largely financial.

Remember, the aim of this methodology is to ruthlessly expose the core processes that determine web success or failure. Other things may be going on within the visitor's path -- such as backing up and repeating page views -- but such considerations are irrelevant to Integrated Path Analysis.

The path marketers follow breaks down into three stages:
        • Getting people into the site
        • Holding them while they do whatever they need to prior to commitment
        • Then getting them to commit to the target action

If the aim of the site is to get contact information, then the target action is submitting the contact form.

If the aim of the site is to sell goods, then the target action is submitting the credit card form that completes the transaction.

The following diagram shows the path someone follows from ad to final target action. The numbers indicate our points of measurement in this path. These are where major decision points occur for the visitor.

Major decision points

It is possible to write Integrated Path Analysis in the format of a set of formulas:

This puts online analysis on the same footing as any other form of sales analysis.

The only other important number we need at this point is total visitors. This enables us to convert any of the figures in the Integrated Path Analysis into absolute numbers. Presenting web analysis in this format is instantly understandable by management staffers who are not involved in the web. In addition, for our own purposes, these numbers offer an obvious set of key performance indicators by which we can set targets and assess improvements.

When we do start working on specific improvements, we will need to start using more detailed metrics that reflect activity in those areas.

For example, clickthrough rate tells us something about the design of our ads, whereas average duration for a specific page tells us how people are using that page (see "" for further information in this regard).

However, over the years, I have come to see the information offered by Integrated Path Analysis as the best single statement of affairs about a website we can have. It is viewable at a glance, yet the individual components contain a wealth of information about the performance of the entire online process.

Brandt is an independent web analyst, researcher and academic.  As a web analyst, he specialises in building bespoke (or customised) web analytic reporting systems.  This can range from building a customised report format to creating an...

View full biography


to leave comments.