Web Analytics Strategy: Data In Context

By Rick Allen - Tue, Aug 3-->

-->

AnalyticsConferencesGeneral

I attended the eduWeb Conference last week and was delighted by a reoccurring theme: measured results. The topic of analytics was front and center. During Karine Joly’s opening-session talk on marketing measurement strategy it was highlighted that many people in higher education are using analytics. This is a positive sign. However, an important question to ask is how are people using analytics and is that use effective? Do they simply view broad analytics data, such as visits and pageviews, or do they view meaningful metrics that inform decision-making? During my eduWeb Conference talk, Making Better Decisions with Web Analytics, I discussed the distinction and efficacy of analytics reporting vs. analytics strategy.

Using analytics effectively is a process that aligns your business objectives and website goals, allowing you to track relevant metrics that answer meaningful questions about your website. Analytics strategy puts data in context. Without context, data is meaningless. Consider the value of the following web metrics:

  • 2,000 pageviews
  • 70% returning visitors
  • 80% direct traffic

What does 2,000 pageviews tell you? What action can be taken based on this metric?

Is 70% returning visitors a positive sign? Is this better than 70% new visitors? We all want to attract new visitors, right?

Is 80% direct traffic good or bad? It suggests a strong brand name because people don’t rely on search engines or referrals to find your website. However, it may also mean that your SEO efforts are ineffective or that you need to improve your linking strategy.

Analytics reporting vs. analytics strategy

Analytics reporting is achieved the moment you open your analytics software and look at the dashboard. Analytics strategy is achieved when you have defined business objectives, website goals, and measurement benchmarks (key performance indicators) that allow you to continuously make informed decisions.

Web analytics strategy is determined by:

  1. Business objectives. What is the purpose of your website?
  2. Website goals. What actions do you want people to take on your website to meet your business objectives?
  3. Key performance indicators (KPIs). What relevant web metrics can be used to measure the efficacy of your website goals over time?

When you have defined objectives and goals you are able to ask meaningful questions that can be answered by relevant metrics. The questions you ask should cater to your organization’s needs, but here are sample analytics questions (PDF) that all web teams should be asking (and answering):

  • How do we decide what new content to publish?
  • How do we decide what content to update?
  • How do we decide what technology to use?
  • How do we decide where to target ads?
  • How do we prioritize web projects?
  • How do we determine appropriate communication channels?
  • How do we know if our content is effective?
  • How do we measure success?

I never meant to be an analyst

My principal function is content strategy. Web analytics analysis is not my primary responsibility. (Ideally, every web team would have a dedicated web analyst.) But, like most higher education web professionals I’m charged with making decisions and recommendations about the direction of web strategy and the efficacy of marketing, communications, design, and content. As a result, web analytics strategy needs to be part of my professional portfolio, and likely yours.

I’ve had some great conversations with folks from eduWeb since my analytics talk and would love to bring the discussion online. How do you use analytics at your school and what are the challenges in making data provide valuable insights? Let’s talk.

This post was written by:

Rick Allen

Rick Allen is the Manager of Web Content at the Babson College F.W. Olin Graduate School of Business and the principal of ePublish Media, an independent web publishing consultancy focused on content strategy and user experience design. He is also the founder of Content Strategy New England, a community of web content professionals aiming to bring clear communication to online user experiences through content strategy.

14 Responses to “Web Analytics Strategy: Data In Context”

  1. Says:

    Great post. I like the distinction you draw between simply looking at stats and incorporating stats as a tool for measuring progress toward a defined goal.

    I think these lessons are particularly relevant when it comes to social media, where there is less clarity about what’s effective and what’s not. Is a Page with 2k fans good or bad? What sorts of status updates should a school be posting? What is the value of 1k Twitter followers?

    I think social media will likely require more trial and error. Most schools don’t have a clear sense of where Twitter or Facebook can be effective (Is Twitter useful for recruiting students? soliciting donations? Can a Facebook Fan Page improve turnouts at alumni events? Can a Group improve yield numbers or retention rates?). The more schools define a goal, determine an appropriate metric, and measure progress over time, the better answers they will have about performance of these newer communication tools.

  2. Says:

    If you want to learn more on the state of online analytics in higher education, you should definitely have a look at the survey results we published last week at https://www.higheredanalytics.com - the official website for the Analytics Revolution we want to start in higher education.

    Our plan is to start with benchmarking, because that’s a good first step to get people in the habit of tracking some metrics. Once we have picked their interest, we will get the conversation rolling on how to use insights to change the way marketing decisions are made in higher education.

  3. Says:

    Mark:
    Thanks! I agree, things do get interesting when tracking a web strategy across social media channels. Measuring “engagement” and “influence” is more challenging than measuring “retention” and “reach”. However, the planning process is the same. If you can identify *what* your measurement goal is, followed by pertinent questions, it’s much easier to find meaningful metrics. By itself, 2,000 Twitter followers is no more meaningful than 2,000 unique visitors. If the followers/visitors aren’t the people you want to attract, they won’t support your goals. Context is needed. I think those questions you raise regarding the efficacy of social media are great too. Most people treat their Facebook page like their website, asking “When?” instead of “Why?”

    Karine:
    I think your higher education analytics revolution is a great idea. I’m on board! I look forward to seeing where this goes.

  4. Says:

    Very interesting post, it’s always good to get ideas for how other people are making sense of Analytics.

    I would argue, however, that new vs returning visitors is not necessarily a measure of content quality (on page 3 of the pdf you claim ‘understanding if a visitor is new or returning is a gauge of the quality of the content. If it’s valuable, people will return’).

    Imagine that a person comes in to a deep page via a search engine, and the page they land on answers their query entirely. They won’t necessarily have any need to return to the website, or to even go to any other page. So that person would register as a bounce, and a one-time visitor. In this example that data is a sign of very good quality content.

  5. Says:

    Milly:
    Thank you. You’re absolutely right. Actually, that’s the point of this post. Context is everything. Rarely does any one metric answer a question. It’s important to identify multiple metrics to support your answers. Another common gauge of quality content is “Time on Page” because it suggests visitors are engaged; however, it can also mean visitors are confused and spending time trying to understand *poor* quality content.

  6. Says:

    Great post, Rick! It’s so great to hear all this analytics talk! :)

    I think segmentation needs to be mentioned here as well.

    Segmentation is essential. Standard segments and beyond. What segments tie directly to our goal?

    Maybe we’re running some branding campaigns and want to segment on conversions by branded keywords vs. generic keywords or campaign traffic vs. non-campaign traffic.

    What about segmenting by in-state vs. out-of-state - how do they behave differently? How do visitors from different referrers behave differently?

    Segment conversions by referrer - What referrer is driving the most action (engagement!) or conversion)? Is there some page on our university main website (if applicable) that’s sending us traffic that’s converting 50% more than our other referrers? Why?

    If we’re in the midst of a redesign, what about a mobile segment? Do we average 5 page views per visit for web users and only 1 page view per visit for mobile users?

    • Says:

      Thanks Shelby! I like hearing all this analytics talk too!

      I’m really glad you brought up segmentation. I agree it’s a critical part of this conversation-on equal footing with the *function* of creating goals. Recently, while preparing for our analytics talk at SIM Tech in October, Jess Krywosa convinced me of the need to focus on this topic. So, I’m glad it gets credit here.

      However, I think talking first about segmentation-as well as the *function* of creating goals-is putting the cart before the horse. You offer some great segmentation examples (I particularly like the mobile segment!), but it’s hard to identify that *relevant* data without first determining the questions you want to answer. That’s what I want to convey.

  7. Says:

    These are great tools, i personally use clicktale which has the best of the bunch in my opinion. Its heatmaps are second to none.

  8. Says:

    I agree that we need to do more analysing. I tend to ignore most of what GA throws at me, to be honest. I just can’t think of a way that it can be useful to me.

  9. Says:

    Very true, Rick. You need the fundamentals first - goals, KPIs, targets. I think segmentation is part of the cart, though. I really do. We need to think of it as a fundamental. Can decisions be made on data that hasn’t been segmented? Sure. But those decisions could be (and probably are) based on misleading data.

    • Says:

      Shelby, I think we’re on the same page here. I’m not suggesting decision making before segmentation-I’m talking about *planning* before segmentation. Simply, you should decide what data you need before starting to look for it.

      I’m drawing a distinction between strategy goals (planning) and analytics goals (an analytics tool function). The former is the important first step (ask: what do I need to know?)-the latter can be satisfied in other ways, particularly segmentation (as you illustrated in your examples).

  10. Says:

    Article published in Sphinn: https://sphinn.com/story/155921

  11. Says:

    Rick - totally with you. Have you checked out Matt Smedley’s “web analytics framework” he did for the Market Motive web analytics class? It does exactly what you’re talking about here - building a framework for solid analysis. Starts with business goals >> website goals >> KPIs >> targets for those KPIs. Really great stuff.

    - Full blog post here
    - Graphic here

    • Says:

      Shelby, thanks for sharing that post. Matt’s “Web Analytics Framework” is a really great example of what I’m talking about here. I recommend others check it out.