{"id":383,"date":"2012-10-30T13:18:37","date_gmt":"2012-10-30T20:18:37","guid":{"rendered":"http:\/\/pln.lanecc.net\/webteam\/?p=383"},"modified":"2012-10-30T13:18:37","modified_gmt":"2012-10-30T20:18:37","slug":"where-do-visitors-go","status":"publish","type":"post","link":"https:\/\/blogs.lanecc.edu\/webteam\/2012\/10\/30\/where-do-visitors-go\/","title":{"rendered":"Where do visitors go?"},"content":{"rendered":"

One of our goals for this redesign has been to simplify the website navigation for our\u00a0 visitors. We’re taking a three pronged approach:<\/p>\n

    \n
  1. Reduce the number of pages on the site<\/li>\n
  2. Unify navigation<\/li>\n
  3. Simplify entry page navigation<\/li>\n<\/ol>\n

    The first one is the most rewarding – we’re getting rid of pages. So far, we’ve cut 5920 pages by either deleting or moving them to a separate archive site. Hopefully we’re narrowing the focus of our site and reducing the number of links that would take you away from what you’re looking for.<\/p>\n

    The second two approaches are much harder, and will be the focus of this post. Both require a pretty serious knowledge of where people go our our website, as well as what people are looking for. The first place we’ll look for this data is our own internal search traffic, which we graphed<\/a> a few posts ago<\/a>. That gives us 50,000 searches for 3700 different terms over the last three months. We’ll scale it by a factor of 4 to get an estimate of what traffic looks like over the year.<\/p>\n

    Our second source is much bigger, 1,000,000 searches on Google for 500 different terms. We can immediately throw out the 83% that are people searching for the Lane homepage. Then we can combine similar searches: in addition to obvious misspellings, like “Modle” and “Moodle”, there’s terms like “fall classes” and “spring classes” that are really just searches for “schedule”. It’s important to combine the right things –\u00a0 if we combine too many terms, we’re losing an opportunity to provide a more direct link to a resource. But if we combine too few, then our rankings might not be right.<\/p>\n

    When we’re all done, we can combine the Google searches with the internal searches, which leaves us with 278 different search terms. Here’s the top ten:<\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n\n
    Search Term<\/strong><\/td>\nGoogle<\/strong><\/td>\nInternal<\/strong><\/td>\nTotal<\/strong><\/td>\n<\/tr>\n
    staff directory<\/td>\n8569<\/td>\n7832<\/td>\n16401<\/td>\n<\/tr>\n
    mylane<\/td>\n14196<\/td>\n1520<\/td>\n15716<\/td>\n<\/tr>\n
    moodle<\/td>\n12314<\/td>\n1756<\/td>\n14070<\/td>\n<\/tr>\n
    map<\/td>\n7934<\/td>\n5944<\/td>\n13878<\/td>\n<\/tr>\n
    testing<\/td>\n4036<\/td>\n6384<\/td>\n10420<\/td>\n<\/tr>\n
    library<\/td>\n6326<\/td>\n2928<\/td>\n9254<\/td>\n<\/tr>\n
    jobs<\/td>\n6594<\/td>\n2148<\/td>\n8742<\/td>\n<\/tr>\n
    nursing<\/td>\n3403<\/td>\n5252<\/td>\n8655<\/td>\n<\/tr>\n
    downtown center<\/td>\n5791<\/td>\n292<\/td>\n6083<\/td>\n<\/tr>\n
    express lane<\/td>\n2631<\/td>\n3312<\/td>\n5943<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

    Our next step is to look at our actual page views. That’ll let us validate our search traffic, and tell us where our current navigation works. If a page is getting visited a ton, and no one is searching for it, then we did something right… right?<\/p>\n

    We’ll look at our 500 most popular URLs. Once again, our first step will be to combine them where appropriate, a job made somewhat more difficult since we’ve been switching URLs around as we move into Drupal. We’ll also look out for things that unfairly influence our page counts. For example, some of the computers on campus automatically open to the computer labs landing page – 65% of the page views on that page are “entrances”, meaning it was the first page viewed. To account for this, we’ll subtract the number of entrances, minus the number of times “computer labs” was searched (searches also generate entrances, and we don’t want to discount those). Our page views will now be:<\/p>\n

    Views = Views – (Entrances – Traffic from Search)<\/p><\/blockquote>\n

    Sadly, that’s a real pain to calculate due to how our data is set up. But we have a pretty good idea of where this is applicable, so I’ll hand calculate it on the half dozen pages where it’s needed.<\/p>\n

    When we’re done, here’s our 10 most popular pages:<\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n\n
    Page<\/strong><\/td>\n<\/td>\n<\/td>\nPageviews<\/strong><\/td>\n<\/tr>\n
    \/schedule\/<\/td>\n<\/td>\n<\/td>\n1560548<\/td>\n<\/tr>\n
    \/collegecatalog\/<\/td>\n<\/td>\n<\/td>\n196972<\/td>\n<\/tr>\n
    \/library\/<\/td>\n<\/td>\n<\/td>\n148587<\/td>\n<\/tr>\n
    \/explanestaff\/<\/td>\n<\/td>\n<\/td>\n144550<\/td>\n<\/tr>\n
    \/calendars\/academic.html<\/td>\n<\/td>\n<\/td>\n112589<\/td>\n<\/tr>\n
    \/counseling\/steps-new-students-enroll-credit-classes<\/td>\n<\/td>\n<\/td>\n94353<\/td>\n<\/tr>\n
    \/collegecatalog\/career-and-technical-programs<\/td>\n<\/td>\n<\/td>\n93282<\/td>\n<\/tr>\n
    \/testing\/placementtest.htm<\/td>\n<\/td>\n<\/td>\n91167<\/td>\n<\/tr>\n
    \/esfs\/admissions<\/td>\n<\/td>\n<\/td>\n89131<\/td>\n<\/tr>\n
    \/es\/students.html<\/td>\n<\/td>\n<\/td>\n83761<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

    I’d rather use unique pageviews, to see what pages are important to people, rather than the pages visited most, but unique pageviews are impossible to accurately combine. But we’re really just doing a back of the envelope calculation here, just finding a starting point for our navigation. We’ll perfect it as we look at how people interact with the website after the launch of the new design.<\/p>\n

    At a quick glance, we see some differences between here and the search data set. But that’s somewhat explained by looking at the pages people go to while applying for admission – the steps to enroll and \/esfs\/admissions. It seems people don’t search for the application, they search for the institution, then apply. But overall, there’s a good deal of overlap in our two data sets. That’s good – it means our search data is a fairly valid representation of what people look for on our website.<\/p>\n

    As part of the new design, we’ll be adding landing pages for each of our visitor types. Students, Employees, Future Students, and Community Members will all have a page full of relevant links and information. Since each of those pages will be linked from the homepage, we can create a goal: within two clicks of the homepage you should be able to access all of the pages that account for at least 85% of our page views. Put another way, if you land on the Lane homepage, there should be an 85% chance of getting you where you want to be with only two clicks of the mouse.<\/p>\n

    In addition to the landing pages, we’ll also have a dedicated Mega Menu on all of our pages on all of our sites. Since there isn’t as much real estate to work with, we can’t have quite as many links. But we’ll also need to use all of the data we’ve collected to build these links with the same goal – the pages people need should be easily accessible.<\/p>\n

    Unfortunately, now we’ve reached the point where things get a little more subjective.<\/p>\n