How to Avoid Gridlock in Web Traffic Analysis

You’ve launched an online campaign. Now how do you measure its success? Despite all the data that can be sliced, diced, scrutinized and parsed, the process of tracking what’s working — and what isn’t — doesn’t have to be overly complicated.

If you set a clear, reasonable objective and have a system in place to measure a campaign’s success, the process doesn’t need to be a daunting one.

Analytics Tips

Here are seven tips for getting the most out of your analytics — without getting bogged down in the minutiae:

ol.thisol { font-weight:bold } ol.thisol span {font-weight:normal }

1. Develop a “success metric.” The “success metric” is the desired action that takes place when a visitor arrives at your landing page or Web site. Success metrics all invariably fall along the “conversion continuum.” The continuum starts with advertisers whose goal is to simply generate a site visitor and it ends with advertisers whose goal is to complete a transaction and capture payment on that visit. In between, there are advertisers who have “soft” offers, where there is some desired action on the part of the site visitor, but not a credit card transaction.
2. Establish an allowable cost. You’ve established your success metric, now you need to understand what that action is worth to you. There is a direct correlation between where a success metric lies on the conversion continuum and what the cost for that conversion will be.

   Many advertisers do not have a handle on their allowable marketing cost for achievement of their online success metric. A discussion around allowable cost is one of the first conversations you should have with the decision-makers at your company. Understanding your allowable cost will help you optimize your campaign, and help you set more realistic expectations if you don’t believe you can achieve allowable cost goals.

3. Gather the right data.> Employ a promotion code or similar method to measure the traffic that arrives at your site and track it as it proceeds through your conversion funnel to the completion of their success metric. Each creative version and traffic source should have its own promotion code.
4. One variable is better than two (usually). Unless you have a very sophisticated testing platform that allows you to track and optimize multi-variant tests, limit testing to single variable tests so you can understand which variables are effective and eliminate the ones that are not.
5. Don’t succumb to analysis paralysis. In the days before e-commerce, the challenge for brick-and-mortar retailers used to be gathering enough data to inform their actions. They’d rely on interviews with patrons during store intercepts to build a data set. Today, the opposite challenge is true: There’s so much data that it’s easy to get overwhelmed. You can literally measure everything about how consumers interact with your site right down to the individual user level. But do you need to? Limit data collection to what’s actionable and limit testing scenarios to what’s measurable.

   For many campaigns, Google Analytics offers everything you’ll need at no cost, and they do a pretty good job of presenting data in a user-friendly format. For more complex analysis, products like Omniture’s SiteCatalyst and Web Trends offer sophisticated measurement, analysis and optimization tools for online marketers (at a price).

6. Find partners that support your analytics efforts. Most online advertising networks limit the data they provide to their advertisers. Look for companies that pass the most critical data back to you — either raw or analyzed — to help you optimize your campaign. Getting the whole picture lets you hone in on where you’re most successful, and where your efforts are falling flat.
7. Test, measure, validate — repeat. Online advertisers can learn a lot from their direct mail marketing counterparts when it comes to testing. The long established rule of direct mail is to test to establish a control. Validate that your control results are repeatable and test against your control until you can beat it. Online marketers would do well to embrace this paradigm.