Advanced Web Metrics

Key Performance Indicators (KPIs) are used throughout organisations for defining success. They are particularly essential in web analytics due to the plethora of data collected. In fact without KPIs, it is easy to become overwhelmed. So once you have set your overall web site objectives, use KPIs as the metrics to benchmark your progress. By definition, these are a small subset of “key” information points taken from your web analytics reports.

A prerequisite for benchmarking is having KPIs that are in context and temporal. So for example, saying “we receive 10,500 visitors” is a piece of underlying data that raises more questions than answers. However saying “our new visitor acquisition was up 10% week on week” is a powerful KPI placing the metric in context (new visitors have increased) and is temporal (over the past 7 days).

While most KPIs will be averages, ratios or percentages for this very purpose, it is sometimes more insightful to use raw numbers. Why? Because the primary purpose of a KPI is to drive change. So if your KPI conveys the message more clearly and has more impact as a raw number, use it. A good example of this is when money is involved. Everyone understands $$$ and this is always a powerful rationale to change something. However, always keep your KPIs temporal.

Even if your web site is a non-transactional site i.e. you don’t directly sell anything from it, you should always monetise your goals. This enables you to estimate what your web site is worth but also put into context how much it costs you to maintain it.

Example raw number KPIs include:

• Our web site lost 15 orders yesterday because our e-commerce server was down for 34 minutes
• We lost $10,000 in potential revenue last week because our booking system does not work for visitors that use Firefox
• We spent $36,000 last month on PPC keywords that did not convert.

Clearly knowing if any of these numbers are increasing/decreasing as a percentage, or what fraction of the total they represent is important. However, the impact of these raw numbers, particularly when in monetary terms, is far greater at obtaining action and therefore should be the KPI.

What KPIs drive change in your business?

Beyond the obvious use of sales data, form submission and banner click throughs, most organisations struggle with defining their KPIs. Those listed are clearly imporant metrics but they are also very black and white - a visitor either converts or not and so this tells you nothing about how close that visitor came to converting or what level of engagement they achieved whilst on your web site. Therefore, what other KPIs do you consider the most effective at driving change within your organisation?

I have heard the notion of Google being analogous to a bank for a number of years. Recently, Jim Sterne also referred to this bank analogy while we were discussing online privacy at the Orion Analytics panel of SES London. So I wanted share and expand upon this discussion.

Please take a moment to read my disclaimer before continuing - that is, the views express on this site are entirely my own and do not represent those of my employer.

Is Google entering into online banking?

In this respect no (I am not considering Checkout here). What I mean by being analogous to a bank, is in the way that data itself has become “currency”. Information has always been valuable - no one likes to be the last to know, and being the first to know gives you a competitive advantage. So whether online or not, the storage and access of data, its security and privacy, is of paramount importance to anyone who hands over information to be stored by a third party.

That is pretty much everybody with Internet access and includes individuals, small businesses and corporations alike. Google occupying such a strong online position - the conduit for many people to find information, products, businesses etc. obviously has an important role to play when it come to data privacy and security.

It’s important that privacy and security have no intrinsic value

A key factor that makes a good, secure and private data storage system, is the strict application of a one size fits all privacy and security policy. That is, regardless of whether you are one of the worlds largest brands spending millions of dollars on operations, manufacturing or marketing, or a small company selling your handicrafts locally - your data should be treated with equal importance.

As an example, imagine if a search engine or bank provided a greater level of security and privacy to large advertisers rather than smaller advertisers (or to those that saved more than others). Even the perception of such a practice would be disastrous for the organisation in charge of protecting your data “currency”. You of course want bullet proof controls whether you save $1 or $100 million dollars. That is how the banking system works and is also how Google operates with respect to handling all visitor and customer data.

» Read the rest of this entry »

This started off as a reply to a comment from Sara Andersson (of search-input.com) concerning my post about the accuracy limitations of web analytics and the difficulty of aligning data from disparate sources: 2008/02/16/accuracy-whitepaper/#comment-2153

However, the subject is broad enough to warrant a separate post, and probably subsequent ones too! To start off, I paraphrase Sara:

“Most sites are not conducting e-commerce and we [marketers] need to spend our resources analysing the lead generations through a combination of the online traffic to the other trend tools available. In your opinion, how would you best go about to study a trend when you have to aggregate additional resources of information from various tools - all with their accuracy problems?“

As I said in my initial reply, this is a very good and difficult question and it got me thinking…

Huge sums of money have been invested by many a corporation in trying to achieve this, but the fact is, its difficult. For example, how can you compare a TV audience that is passively digesting your ad which mentions your web site (with an unquantifiable number of viewers reading a newspaper, checking emails, making a cup of coffee, taking a bathroom break, putting the children to bed or any other activity that takes place during an advertisement break), with a pro-active person that has connected to the web, conducted a search engine query with specific keywords, and landed on your web site?

Assuming some TV viewers remember your brand or web URL and later visit your web site, are these two types of visitors comparable? The simple answer is NO.

Marketers still need to analyse and compare data from disparate sources

So what can be accomplished? First, lets set up some examples:

Example 1: You are a non-eCommerce web site and you are using and ad serving platform (e.g. DART from DoubleClick) to display banners on a selection of related portal sites in order to drive enquiries to your web site. Your web site’s call to action is to generate a telephone call to your call centre to complete the transaction. You use an internal tool to track how many calls are being made as a result of this.

Example 2: In addition you are running a print campaign (magazine) driving brand recognition by directing visitors to your web site and you wish to measure its impact. All you have from the magazine publishers, is a monthly distribution figure of 50,000 readers.

How to integrate data

The key to be able to make comparisons is to ensure you have an integrated approach to your marketing That is, don’t treat different channels as silos. Keeping web analytics at the heart of your strategy gives you the best possible chance of success - consider Figure 1. Put simply, the web is the most accurate form of marketing measurement. Therefore use it to your advantage.

Using Example 1, integration could be achieved by the use of a unique telephone number on your web site landing pages. Ensuring your specific DART landing pages are not accessible by non-DART referred campaigns (not linked to within your web site or visible to the search engines), your call centre could then identify these calls as originating from a particular banner advertisement.

In fact, a clever CRM system can dynamically assign the page’s unique telephone number on-the-fly depending on the referring source. For example, if from a DART banner use: 1234-5678, if not use: 4321-09876. By this method, you are able to directly measure, and differentiate, how many visitors and how many telephone enquires are received from your DART campaign.

» Read the rest of this entry »

Its been a long time since my last post (October!) but I have still been busy writing - essentially finishing the edits to the book and that has given me plenty of ideas for future posts… Anyhow, happy new year!

This post is an extrapolation of a section in Chapter 2 of the book which has led to a separate PDF accuracy whitepaper. Why is this necessary? Well, the question of accuracy crops up all the time on numerous forums and at conferences. Essentially many practitioners of web analytics worry about accuracy. Some vendors even claim greater accuracy than others (though as I explain in the whitepaper this cannot be true), and there is the inter-industry debate about whether off-site analytics (for example, Hitwise, comScore, Neilsen//Netratings etc.), are better at predicting traffic levels than on-site analytics tools (such as Webtrends, Omniture, IndexTools, Google Analytics etc.). I won’t go into that debate here, except to schematically illustrate the two different web analytics approaches in Figure 1.

Figure 1: On-site v off-site web analytics

The truth is, for either approach, web analytics is not 100% accurate, and even the error bars are difficult to measure. But just how important is this?

The initial reaction to this question is usually “very important”. Marketers need to manage visitor acquisition budgets, content creators need to be know if their work is engaging (building relationships or not), e-commerce managers need to know their conversion rates, and webmasters/developers require data on which to base decisions for technology investment (internal site search, rich media applications, checkout systems). But is this a question of accuracy or is precision the more important factor?

Wikipedia has an excellent article on the difference between accuracy vs precision and I am reliably informed that Jim Novo spoke eloquently about this at the 2007 eMetrics Summit in Washington (though I wasn’t present!). From the Wikipedia article, the two targets shown in Figure 2, schematically illustrate the difference beautifully.

» Read the rest of this entry »

When viewing Google Analytics reports, I constantly need to remind myself of the difference between goals and conversions (may be its just me that gets confused..!). Whatever, I thought I would share my clarification.

One obvious difference is that a transaction is associated with an e-commerce completion (a purchase) while a goal conversion is considered a non-ecommerce conversion, such as a PDF download, a form completion or visit to a special offers page. Of course they are all conversions, which is where I think confusion lies.

Explanation: The most important difference as far as Google Analytics is concerned, is that a conversion can only happen once during a visitor session - that is, a visitor can only become a customer (convert) once and that makes sense. However, if one of your goals is set to *.pdf for example (any PDF file download), then should a visitor download 5 PDF files during their session, it will only show as one conversion in your Goal Conversion reports. Assuming you are tracking pdf downloads, the 5 PDF files will of course show in your Content > Top Content report.

» Read the rest of this entry »