by Dorota Umeno on February 06, 2011
You’ll hear us talk about data, metrics and measurement all the time around the Confluence office. We’re unabashed geeks. One of our favorite quotes is: “If you can’t measure it, you can’t manage it”. It is variously (and credibly) attributed to Peter Drucker, W. Edwards Deming and four or five other notorious individuals.
Even the juiciest web analytics data is just a bunch of numbers, meaningless unless it contributes to key performance indicators (KPIs). KPIs are measures that translate raw data about visitor activity on a site into actionable business intelligence. KPIs function as trackable benchmarks that enable decision makers to evaluate the performance of a site over time as well as define, manage and track progress towards business goals.
If you haven’t thought about this, you’re not alone. Many companies collect analytics data about their site without giving it much thought. So, inspired by the excellent online WAA-UBC web analytics course we completed a few months ago, we thought we’d provide some basic guidelines on what constitutes a KPI and how to figure out which measures to pick for your website.
To determine what KPI is most relevant for a website you first need to know what is the main purpose (goal) of the website. This can get tricky since many (most) websites can have more than one goal, for e.g. driving visitors towards online sales (eCommerce) or to fill out a contact form for the purpose of lead generation. Success would be defined in different terms for these two activities. Luckily analytics allows for data segmentation, so it’s a bit more complicated, but not impossible to define KPIs for one part of the site and track them separately from a different part of the site.
In general most websites fulfill one of the following four purposes:
The purpose of these sites is to get visitors to complete an online purchase. ECommerce sites come in all shapes and sizes from mom-and-pop shops with a digital storefront to the 800 lb. Gorilla, Amazon.com. The purpose of an eCommerce site is to generate revenue, pure and simple.
Top KPIs for eCommerce sites include:
(1) Order conversion rate – the ratio of visitors to completed orders.
(2) Checkout conversion rate – the ratio of initiated checkouts to completed orders.
(3) Average order value (AOV) – the amount of the average purchase.
(4) Average visit value – the ratio of orders to site visits.
(5) Customer loyalty – ratio of new to existing customers, and
(6) Search engine referrals – the ratio of referrals from natural and paid search results to the site as benchmarked against industry average(s).
The primary goal of lead generation sites is to get visitors to call the sales department, submit a sign up form, subscribe to an email list or newsletter or otherwise leave contact information enabling a follow-up call. Businesses that do this typically sell products or services that require careful consideration and represent a significant investment of money or long-term customer commitment. Those types of transactions are information intensive, rely on trust and typically require personal contact with a sales representative for purchase completion. Insurance, investment and mortgage companies are examples of such sites.
Top KPIs for Lead Gen sites include:
(1) Lead conversion rate: the ratio of site visitors to leads generated.
(2) Cost-per-lead (CPL): the ratio of marketing costs to total leads generated.
(3) Single access ratio (bounce rate: the ratio of single page visits, where a visitor only looks at the first page they access (aka landing page or entry page, quite often the home page) to total number of pages visited for the site.
(4) Traffic concentration (page popularity): the ratio of visitors to a page or section of the site to total number of site visitors.
The purpose of content sites is to attract visitors and keep them engaged for as long as possible. They provide interesting, engaging and regularly updated content, often in the form of expert articles or media, which is provided “free of charge” to the public and being supported by advertisers whose ads are displayed to the visitors as they read and browse the site. Some content sites function as corollaries to offline content, for e.g. elaborating on magazine articles or television programming to drive sales and other offline purchases. Examples of content sites include ESPN.com and NewYorkTimes.com. As such content sites rely on fresh, engaging content that has visitors coming back again and again and spending time on the site to attract ad revenue to the site and the offline property as well.
The most relevant KPIs for content sites include:
(1) Depth of visit: the ratio of page views to visits.
(2) Returning visitors: the number of unique visitors who visited the site at least once before. This is a tricky metric since some 30% of people regularly clear their cookies, removing our ability to identify returning visitors as such – they get re-recorded as “first time visitors” again.
(3) New visitor percentage: the ratio of new visitors to unique visitors.
(4) Page/content depth: the ratio of page views to visitors at the page or content level. It is a good idea to measure strategically important (key) content separately to observe changes in engagement over time.
(5) Visit duration: how long visitors stay on a specific page or on the site as a whole.
Customer service and support sites typically have two goals: 1) increase customer satisfaction and 2) reduce cost of in-person customer service at call centers by reducing the need for direct interaction with customers requiring help. Customer service sites follow a self-service format. They rely on high usability to enable visitors to quickly and efficiently find a solution to their problem or an answer to their question about a product or service.
KPIs for measuring effectiveness of customer service sites include:
(1) Customer satisfaction metrics. Those are typically collected through requests for product ratings or satisfaction surveys.
(2) Time on site. As with content sites, it matters, but whereas content sites want visitors to linger longer, too much time on site in this case may mean that visitors are not finding what they are looking for.
(3) Percentage of visits of fewer than 90 seconds. A subset of the overall time on site metric, if it is high it could be an indicator of bad usability (we will elaborate on the interpretation of the measures in a follow-up post).
(4) Content depth: in contrast to media sites, the goal should be to make it quick and easy for a user to find the information they are looking for or to perform the self-service task.
(5) Top on-site searches: not a traditional KPI but important in identifying areas that may require better navigation or more content to meet customer demands.
Not really their own category, hybrid sites – as their name implies – represent a mix of the four main types I discussed. Two prominent types of hybrid sites that emerged in recent years include social sites like Facebook and wiki sites like About.com. Those types of sites combine elements of support sites with rich content, supported by advertising. They include crowdsourced or community-created content and they encourage engagement to attract advertising dollars. The appropriate KPIs for such sites would be ones that work for those two site categories. However, it is not as likely for a small business to develop this type of site. It seems that this category firmly belongs to a few big players.
The bottom line is that the KPIs you pick may represent a combination that makes sense for you based on the nature and goals of your business and the type of site you create to generate revenue for your business. Measurement is only of value if you measure what matters to you.
Stay tuned for our next post in our Mini Primer series in which we will talk about calculating KPIs and elaborate on their interpretation for the purpose of website optimization.
Contact us if you want to chat about metrics, measures, KPIs or have questions about any other aspect of your digital strategy. We’re always happy to answer questions.
Check out these related posts: