Get Agile with Web Analytics: Part 1
Time is, of course, money. If you really want to see an ROI from web analytics, you need to reduce the time to insight. This may mean slaughtering some sacred cows and taking a fresh look at what you’re tracking and why. In this series, I outline ways to abandon bureaucracy and obligatory reports, take action faster, and point out why more automation is not always the answer.
First Things First
As with any other initiative, if you don’t know where you’re going, you’ll never know if you arrived. Save massive amounts of time and cost by ensuring your Key Performance Indicators (KPIs) are clearly defined and accepted throughout the organization. For most websites these can include:
- Purchases or revenue goals
- Generated leads or quote requests
- Acquisition rates for key target audiences
- Email signups
- Trial downloads
- Dealer contacts
- Site engagement levels
- Ad impressions
Elevate your KPIs in your reports and focus on outcomes. In most cases, your web analytics tool (once configured correctly) will automate the tracking of these metrics (also providing your data gaps are closed – see below). If this is challenging due to the use of numerous third-party systems, you may want to manually log performance over time in a spreadsheet or work toward automating KPI data collection via APIs. These are also good candidates to be featured in custom-designed reports for the top brass in your organization.
Close Those Data Gaps!
Now that you’re in agreement as to why your site exists in the first place, it’s time to make sure you can actually verify if it’s achieving set objectives. In some cases, your site may not be configured in a way that allows for the right data to be collected at all. Minor site updates can have major impact. Examples include:
- Contact forms don’t submit to a unique URL making it difficult to track conversions in your web analytics tool
- Single pages contain content for multiple topics: Breaking topics out into separate pages allow for individual pageview measurement, segmentation, and targeted landing pages.
- No site search! Without a search function on your site, not only are you depriving users of a helpful tool, you’re also missing out on seeing the terms they enter into your search box. Learn what your customers are NOT finding by reviewing internal search data and the start pages that trigger the most queries.
- No campaign parameters: You advertise a key promotion from various places (homepage, banner ads, external sites, email), but lacking proper campaign parameters in your URLs, you can’t tell which source is performing best. (More on campaign tracking in Google Analytics)
- You hand off traffic to affiliates or partner sites without measurement: For example, your main site features product overview information but the actual purchase takes place on a partner ecommerce site. If sharing clickstream data across two different profiles isn’t possible, at least track referrals to the partner sites as events.
- No voice-of-customer tools: By tracking only the quantitative, you’re only getting half the story. Add rating and comment tools, opt-in surveys, and social sharing tools to collect and compare the qualitative response.
- Cryptic URLs: If your URLs are full of parameters and session IDs, or worse, are dynamically created based on various scenarios, it will be difficult to sort out which page is which in your analytics data. Work to standardize URLs, file names, and page titles.
Create Your “To Review” List
In the early days of web analytics in order to gain insight, data needed to be pulled out of the web analytics tools into spreadsheets with various custom calculations applied. These days, even free tools like Google Analytics have evolved to the point where many of those calculations are handled for you. You can spend more time generating insight, and less time processing your data.
Even with these significant improvements, one still needs to have a very clear agenda of what to track on a regular basis. This need not be complex. It could be a simple list in a Word doc with approval from all key stakeholders, an Excel spreadsheet designed to be a living document (Google Docs work well for this), or a custom-built web application. The key is, keep the focus on strategy and analysis, not on automation and reporting.
Your agenda will likely include some of the following:
- Google Analytics Intelligence: A fantastic, and relatively new addition to Google Analytics that generates automatic and custom alerts of significant events within your data.
- Overall goals and conversions per audience segment (provided you’ve closed your data gaps)
- Campaign performance (again with the data gaps)
- Email campaign performance
- Search marketing performance (both organic and paid)
- Top traffic sources by key segments and audiences
- Specific referring site traffic (who is sending you love)
- Regional performance (where are your conversions and quality traffic occurring)
- Internal search usage (what are your visitors looking for)
- Social media monitoring
Under each of the above you’ll break out the specific campaigns, keyword groups, referring sites, and so on that contribute to success in each area. You’ll also want to group specific metrics by business goal and by key audiences. Depending on your marketing focus in a given month or quarter, you’ll know where your efforts have been directed. It’s in those areas where you may choose to document your performance in your spreadsheet or custom web application. However, in most cases, your web analytics tool (if configured properly) is already doing this for you and you can easily view trends.
You’re ahead of the game simply because you knew, very specifically, which reports to review rather than meandering aimlessly. Save time by documenting (or automating) only what will directly help you gauge whether you’re moving the needles you’re trying to move in a given time period. Take a serendipitous approach with the rest of your list. For instance, if a new referring site surfaces which is sending quality traffic, make a note of it and look for it next month. If all looks the same as last month, don’t bother reporting it. Make your metrics work hard to be logged and apply the “So What” filter mercilessly.
Tune In Next Time
In Part 2, we’ll continue our exploration of an agile web analytics philosophy and cover topics such as data automation (and why it’s not the end-all-be-all), reporting strategies, analysis and insight brainstorming sessions, and the relative importance of data accuracy.
As always, comments are encouraged. What data gaps are you working to close? What’s on your “To Review” list?