Non-transactional Analytics, Part 2

Last time, we took an initial look at the differing needs of transactional and non-transactional sites when it comes to tracking and measuring success. The absence of strongly defined conversion points makes it harder for owners of predominantly informational sites (like may company portals) to understand their activities’ effectiveness and whether they’re getting a return on their investment. This time round, I’ll share some thoughts about ways to measure non-transactional sites.

When it comes to measuring non-transactional sites, context is really important. It’s difficult to measure things in a vacuum; you need something on which to hang the metrics or the analysis. On a transactional site, the context is provided by the transaction. All the focus is around the path to purchase.

In non-transactional sites, that key end point may not be there, but some context is still needed. At a macro level, the context will come from the site’s roles and objectives. Measure to understand overall site effectiveness. At a micro level, the context may come from a specific site page and what it’s there to do. Measure to understand individual site elements’ effectiveness.

People at organizations with non-transactional Web sites tend to have a one-size-fits-all approach to measuring their sites. It’s usually with a Web analytics tool. They’re frustrated because they feel unable to get any insight into what’s working or not working on the site. Often, they use bland metrics such as average visit duration to measure visitor engagement, but the only context they have for this metric is how it changes over time. Quite often, it doesn’t. The problem with metrics such as average visit duration and average pages per visit is they measure across different groups of people trying to do different things on the site. As a result, they’re often meaningless and dangerous.

At the macro level, we’re concerned about whether the visitor interaction was valuable to the visitor and whether it delivered in a way that was valuable to the organization. There’s a key role for surveying visitors or customers on a regular, if not frequent, basis. Visitor satisfaction is a key metric for non-transactional sites. Other key metrics include likelihood to return and propensity to recommend the site to others. If visitors are satisfied with their visit, they’re likely to return and recommend it to others. The site must be doing a good job for them. You can support these metrics by understanding visitor satisfaction with different aspects of the experience or the site, such as:

• Content breadth
• Content quality and usefulness
• Ease of use of the site (navigation, search, etc.)

Now we know if visitors are happy, but what about the organization? Does it serve up this value efficiently and effectively? At the macro level, transactional sites would look at metrics like revenue or profit. On non-transactional sites, there isn’t any revenue, only costs. So understanding the cost to the organization relative to the value served is important. Revenue tends to be relative simple to measure, whereas costs can be quite tricky. There may be direct and indirect costs, fixed and variable costs, all sorts of costs. Quite often, the challenge can be determining what the true, appropriate costs of running the Web site are. They may include:

• Marketing
• Content (the costs of the people generating the content)
• Overhead (the cost of the rest of the team)
• Technology (hardware, software, etc.)

The ability to collect and track these costs varies from organization to organization. If the full costs can’t be easily allocated and pulled together, a consistent approach over time is needed. Cost data can then be combined with other data to understand value relative to the costs. You might track cost per visit or cost per page served to see if the site can scale in an affordable way. You could even combine cost data with satisfaction data to understand the cost per satisfied visitor over time.

You may have noticed there hasn’t been much mention of such metrics as visits and page views as I’ve been discussing these macro-level success metrics. That’s because Web analytics metrics tend to work best at the micro level for non-transactional sites. These metrics need context, and that context may come from comparing different parts of the site or comparing different visitor segments and how they behave.

This is something I’ll look at in more detail in part three.

Till then…