June 2006


29 Jun 2006 12:14 am

pine1I recently had the privilege of doing an interview with James Maguire of ecommerce-guide.com. The overall scope of the interview was to talk about what do small and mid-sized businesses need to do in order to extract value from their web analytics implementations. The goal was to provide recommendations and tips on specific reports and touch on the “hot” area of Search.

You can find the complete interview on the ecommerce-guide website: How to Use Web Analytics, Part 1. It is a brisk ten minute read, I recommend it (of course you expect me to say that about my own interview!!).

Here is a summary of the interview (but not too many details, don’t want to steal the thunder from James).

Overall Theme:

  • Web Analytics is very hard for all of us. Start simple, start free, focus on “Inputs” and “Outputs” in terms of initial reporting efforts and fall in love with SEO.

Getting Started:

  • You don’t have to buy expensive software to start your web analytics efforts. With the complexity of high end packages it pays to learn first using some of the free packages. ClickTracks Appetizer for log file parsing, Google Analytics for Javascript tags (whatever works best for each SMB). Implement free, dive into data, learn through grunt work and move up the food chain to the expensive packages if these are limiting or as your sophistication grows.

    [UPDATE: ClickTracks Appetizer is no longer available. There are many web log parsers available, please do a quick Google search. Alternatively you can start with the lite version of WebLog Expert.]

    Omniture, WebTrends, HBX, CoreMetrics etc are all great packages but I recommend the path above so you will make more intelligent decisions, apply the 10 / 90 rule and extract maximum value.

  • Don’t open your new package and dive in and get distracted. First figure out your desired outcomes. Why does your website exist? What are the goals of your website? Once you have your goals only focus on reports, initially, that will help you understand the website behavior that is tied to those goals/outcomes.

Six Initial Reports I Recommend (to avoid death by data):

  1. Top Key Phrases from Search Engines. Search is the game baby, and even more so for our SMB friends who have very small ad budgets and have to leverage the power of search engines. Do this to infer intent of your customers, do this to know if you are getting the traffic you think you deserve. If you are not take action.
  2. Top Referring URL’s. Taking the above recommendation to the next level. Many SMB’s create partnerships with other SMB’s or associations etc. As you look at the referring URL’s do you see any that are interesting and surprising? Do you have some relationships with them? If not should you? Then, my favorite, segment out those referring url’s that are sending you traffic that meets your goal criteria (a order, a lead, a particular page view etc).
  3. Site Content Popularity: Websites are small and websites are big. Do you know what content is being consumed the most by your visitors? It is probably less than 20%, it is important to know which 20%. In my experience we are always surprised by what content is being consumed by our customers, specially if you segment out content consumed by those who meet your site goals. Start with most popular pages viewed on the site, go from there.
  4. % of Visitors Who Visit the Home Page: This recommendation is more a reaction to the fact that 90% of website owners are way to enamoured by the home page and perfecting it. In reality around half or less of the site traffic sees the home page. Do you know what this number is for you? Knowing this will mean you can assign resources optimally, focus on other pages, if you have “golden content / promotions” you won’t just show it on the home page.
  5. Click Density / Site Overlay: It takes a lot to interpret the Click Density report but there is nothing simpler for a SMB owner to start with. Web Analytics comes to life, they can “see” the clicks and relate to visitors in a new and more profound way. Click Density’s real power comes when you look at segmented click density, where do purchasers click vs everyone else. Actionable insights. For a SMB owner, if you see “interesting” click density behavior do simple experiments with layout, content, navigation to optimize.
  6. Site Bounce Rate: I define this as visitors who stay on the site for less than 10 seconds (or one page only, though for reasons I’ll blog about some day I prefer the 10 seconds). Specially for a SMB owner who has precious few resources to spare each visitor is valuable and it is hyper critical to know what this number is (and then segment by campaigns, referring urls, top paged on site with high bounce rate etc to know who is sending you valuable traffic).

Search Engine Optimization & PPC Point of View:

 For better or for worse Search dominates our lives. A start to effective search strategy is to know if you are optimized for the right key phrases, and all aspects of that:

  • Do you know what the most important key phrases are for you (both brand and category, for a SMB category becomes even more important)?
  • Are you showing up in search engines?
  • Do you have the right meta tags?
  • Do you have the right content with the right key phrases?

Something like HitWise is quite expensive for SMB’s but there are options like the Overture Keyword Selector Tool or the options at MSN adCenter Labs or a very cost effective option such as Wordtracker (one year subscription for $255). Or my cheeky suggestion, go to your direct competitors pages and look for the meta tags by doing a View Source.

In terms of Search Engine Optimization (SEO) and Pay Per Click (PPC) I am quite biased towards a solid SEO strategy first, otherwise you are just renting traffic that will come only as long as you spend money.

I also recommend using a service like XML-Sitemaps to create your site maps and uploading those into Google to get a higher chance that Google will index all your content effectively. XML-Sitemaps is free, Google also does not charge anything to upload so something quick that can be helpful.

So this was longer than I expected to blog about. : ) The interview article is much longer so if you want greater detail and ideas please click here to read it.

There is some high praise for James in this blog post Can’t Live Without Web Analytics for the interview. Check it out if you have a second, I don’t think you can resist a blog post with that title!

Please share your feedback on this blog post / interview. Would you have made the same recommendations for important reports? Are there others that a SMB owner would find more useful? Do you recommend other strategies? Please comment.

[Like this post? For more posts like this please click here.]

26 Jun 2006 12:34 am

purpleAll of us are doing SEM/PPC, many of us are doing SEO but very few of us are paying any attention to our internal site search. Even companies that are extremely sophisticated with search campaigns treat internal site search as a step child / orphan / use your favorite metaphor.

Internal site search is someone visiting your website and using the search feature on your website to find information.

As websites have become complex (just look at amazon’s home page) an increasing amount of site users are jumping to the site search box and using that to find what they want. Internal site search usage numbers are hard to come by but in a simple google search for metrics seems to indicate around 10% as typical (YMMV).

So if 10% or more a site traffic is using your internal site search why don't we pay attention to it?

  • The magnitude of site search usage is usually not a number that is easily available because many of our analytics tools don’t have a way to report them off the bat.
  • There are all kinds of hairy implementations of internal site search software that compounds the ability integrate reporting.
  • There is a wrongly held belief that if you optimize for external search then you are also optimizing for internal search.
  • The true value of this data, and its multiple usage scenario, is not completely understood.

So why should you care about searches on your website?

  1. External searches have nothing to do with internal searches. Here are the external search engine key phrases for this little blog (implication: numbers = small for blog, bigger for your website I am sure) :

    ak external

    And here are the searches conducted for the same time period on the blog itself :

    ak internal

    You will see that while the numbers are tiny there is nothing in common between the two sets of keywords because visitor intentions are radically different. 

    The key insight is that most searchers are looking for generic things (think for example "category terms") in google or yahoo or msn to locate a relevant site. But once they are on the website they are looking for something specific.

    Perhaps you are skeptical that this blog example is not truly "relate-able" to your business website, having done this now for more than 50 different websites I assure you that you will find this to be true for your website.

  2. One immediate implication of the above is that if you are only doing SEO (meta tags, key word creation, url structures, best bets etc) for external key phrases then you are not solving for visitors on your website. It is likely that when people search on your website they will get sub optimal search results.

  3. If you use Click Density / Site Overlay you already know that it is difficult to find a feature (or even the dreaded “path”) that is used by 10% of your site traffic. Yet internal search is a feature that is being used by 10% or more of your site traffic. This is a gift from God for a site owner because you just analyze this one things, optimize results / content and you have a huge winner on your hand.

  4. Lastly internal search key phrases are great for two reasons

    #1) I love finding visitor intent and internal or external key phrases convey intent and it that is very valuable, study them and try to infer customer intent
    #2) They are a great reflection of what is “broken” (ok that is a bit alarmist) about your site navigation/links, stuff people can’t find easily or content that is completely missing.

         Example:If you have a big honking blinking button that says Subscribe Newsletter and yet that is your top key phrase then you might want to rethink the blinking button.
         Example: If one of the top five key phrases on your internal site search report is Register Product and you don't offer registration then here is your customer demanding that you do.

Next logical question, if you have a great search tool and your analytics tool, like ClickTracks for me (see Disclaimers & Disclosures), makes it possible to analyze internal search in one click then how do you measure success.

A few recommendations and analysis starting points:

  1. Atleast on a weekly basis review the top 25 key phrases from your internal search report (you could look for more, just look for where numbers fall off a cliff, usually after the top 20 – 25). You are looking for interesting and surprising key phrases (to solve for the above two examples, are there key phrases for content that you actually have prominently displayed and key phrases that surprise you i.e. content you don’t have).
  2. ClickTracks has a feature (perhaps others do as well) where you can do click density / site overlay on search results. This is magnificent because now you can see if customers think you are serving up relevant results (as opposed us the proud site owners thinking that).

    Have a goal that for your internal site search should be so good that most of the click density should be clustered on the top five search results links and no one should click Next Page (shoot for the moon I say : )).

  3. If you have the GSA the it will allow you to create synonyms and best bets. Use these features for atleast your top 25 key phrases. Then use the Click Density to see if people are clicking on your best bets, if not optimize.
  4. Measure conversion rate (if you sell something) and customer satisfaction & task completion rates for internal site search users (something like ForeSee/ACSI, see Disclaimers & Disclosures).

    The hypothesis is: relevant search results = faster access to relevant data = higher customer satisfaction and task completion rates = higher conversion = more money = higher bonus for you in your annual employee review.

(Sidebar: If you use the Google Search Appliance, GSA, its native reporting out of the box is sub optimal for many reasons, my recommendation is to integrate GSA searches with your analytics tool, like ClickTracks or Omniture or WebTrends or CoreMetrics or HBX if they allow it.)

I hope to have convinced you that internal site search is important for any website, analysis of the data captured can be very insightful and more importantly a actionable money making venture.

Quick Anecdote: A company had woken up to the fact that internal search was important and their search software was terrible. So they spent xx thousand dollars and bought the spiffy Google Search Appliance (GSA)  and put it on all their sites. They are surprised to find a month later that their customers still complained that site search was bad.

Life Lesson: You can put the greatest tool on earth, Google, on your website but if your websites are sub-optimal (url structures, content key words, missing meta tags and best bets) then all GSA will enable on your website is your customers finding crap much faster than before you implemented the GSA.

Agree? Disagree? Am I tooting this internal search horn too much? Would you like to share your own tips about how to analyze internal search data? Do you know of a tool that is really good at this? Please share your feedback via comments.

[Like this post? For more posts like this please click here.]

Next Page »