Sounds too good to be true? It usually is.

Faceted NavigationJulie Capshaw writes: “Tony from {Name Redacted} says nothing would make him happier than your death!!! He had me convinced that you were wrong about everything related to our internal text search until today. Today we got their new contract and you were right.  They’re slimebags! Our price increased SIX times over what it was this year!!! Now where should we go? I need a new vendor ASAP!!!”

Julie!  Julie!  Julie!

When will you ever learn?

As much as I’d love to out {Name Redacted}, I am too busy engaging in Twitter Flame Wars to get into another kitty litter dust-up this week so you’ll have to suffice with this short post and recommendations in a private email.

Do you like Amazon’s internet text search function?

Two days ago, I was searching for Kevin Hlllstrom’s new book.  (The $95 one that I won’t understand not the $.99 one that I can understand and already bought.)  It’s called Hillstrom’s Digital Profiles.  When I searched on Amazon for “Kevin Hillstrom,” it listed his books (print and Kindle editions) and then books about Indians, the Cold War, the Michigan Adventure Guide, and Watergate.  I was on my phone so I kept scrolling and then I finally got to the new book.

So you tell me, was my search experience good?  Bad?  Indifferent?

I’ve bought every one of Kevin’s books.  Amazon knows this.  Amazon employs wizards and fairies with pixie dust and unicorns AND spends MILLIONS  OF DOLLARS to PERFECT their search.  They know all the books I buy (averagely 10-12 books a week – yes, a week) and they know that I am far more interested in marketing than Michigan.  The best thing about Michigan is @netmeg and she’s so mysterious God only knows if she really lives there or even exists for that matter.  She could be a bot. But I digress…

Amazon uses rainbows AND THE POTS OF GOLD UNDERNEATH THEM, yet, they still don’t do it 100% right.

Am I picking on Amazon?  Absolutely not.  I love them.

But if Amazon can’t perfect their text search – in a simple area like FLIPPING books?  Well, there’s not a lot of hope for the rest of you.

Which brings me to Reason #1 that Tony hates me.  I do not believe any internal text search will ever be perfect (because users are not) and promising a package will be?  Heh.  Hilarious.

Reason #2.  (Don’t worry, I am not going through ALL the reasons.  I have neither the time nor the inclination to write a tome tonight.) Many (not all, but many) text search companies tell people all sorts of bogus “facts” (cough) based on their “years of testing and proprietary research” which roughly translates to their sales and marketing departments making up crap and the lawyers blessing it with “no, we can’t really defend it but nobody sues anyway so don’t worry about it.”

For example, most of the companies will tell you to put your search box in the upper righthand corner of your site because that’s where it works best.

They’re right. That’s not a complete lie – it’s one of the top two places. 

What they OMIT however is that many of them know that putting it in the upper righthand corner often increases the number of searches, which may or may NOT be a good thing for your business.  Usually it’s not.  Depends on how good your data is.  A lot of people leave on search – failed searches and searches that you deem “successful” where the user can’t find anything.  (In other words: UNsuccessful.)

The sneaky thing is that only 1/3 of the users typically leave on the search results page – the rest leave on the 2-3 subsequent pages.

Do the search companies tell you that? Yeah.  Not. Usually.

One of the main reasons why they want people to take advantage of your search is a lot of them have commission deals and get paid per search and/or directed search result page.  Thus, they have a VESTED interest in telling you to put it in the place that’s best for them.

Is this a rant?  No.  It’s a reminder for us all (myself included) – when vendors tell you stuff and give you all their stacks of back-up information, it’s usually for a purpose and it’s often VERY biased. 

Review the sources.  Figure out what the sample size is.  Look to see if there was a control group.  Determine what kind of significance there was to the study.  Find out who paid for the research.  Find out if the participants were compensated.   You get the drill.  

If it sounds too good to be true?  It usually is.

Have a question you’d like answered?  Click here now to submit it for FREE or email me at


  1. Sappho Charney says

    Great post! I’d like to add that one of the problems with the search function is that often the customer doesn’t know exactly what he’s searching for, so he runs a series of guesses to try to get at it. We would help him much more if we could organize our products in logical ways on the site, so that the search function didn’t have to be used as an all-purpose trawler.

    For instance, on Amazon, each author’s works can be sorted in a number of ways, such as the baffling “Relevance,” the pragmatic “Price,” and the misleading “Publication Date.” But if you’re searching for, say, an early Ruth Rendell novel whose name you can’t recall, good luck trying to find it. “Publication Date” includes all the reprints, so that’s totally misleading. Your best bet is to click on the name Ruth Rendell, go to the author page, and sort all the books alphabetically. You’ll still plow through 13 screens because you only get a few titles per page (some of them in foreign languages — hello, Search specialists, how difficult is it to include a language selection feature?), but at least you have a chance of stumbling upon your book this way.

    And this is the search process of one of the very best websites!

    The simple solution would be a line-by-line bibliography, which would take up little screen space and display all the author’s works. But, to quote Amy, the goal of any good website is to be a bookstore, not a library. For all its bells and whistles, Amazon knows this better than anyone . . . which is why random books often appear in searches, such as a 2-volume vampire series in the midst of my Ruth Rendell list. Amazon has watched me buying vampire novels for my teenage daughters for 4 years now, and isn’t about to let me escape the site without viewing its latest offering.

    Maybe we need a new post on the ethics of the search function!

    • says

      Excellent comment Sappho.

      I don’t read fiction but the reprints thing drives all my big-reader-friends bananas.

      The thing that folks don’t seem to realize about search is that it will never be perfect. (Hello Google.) If you want it anywhere close to acceptable, you really need to work it by improving your data and your copy.

      When you write SEO copy, are you impacted by
      thoughts of what people will search for? I mean I know you are because you’re a kick-ass copywriter but if you have any tips on how to do it, it’d be great to add them for other folks who may read this post.

      Thanks again!

  2. Mike McCormick says

    “Tony from {Name Redacted} says nothing would make him happier than your death!!!”

    Perhaps if you were nicer to Tony, he wouldn’t be so mean.

    • says

      Tony’s company has gotten over $3 milllion in business from my referrals in the past 2 years. How much nicer should I be?

      • Mike McCormick says

        You should take a Louisville Slugger to him until he stops being happy at the prospect of your death.

  3. says

    I guess it shouldn’t surprise me that an on-site search vendor might say to a client: “Look, put the search box in the upper right hand corner to make it easier for your visitors to find it.”

    Thanks for reminding us that we need to use that hot corner to help move people to do what we really want them to do, and it’s not necessarily to conduct a search. As with everything, test it. This reminds me of my first marketing job after college and a sage boss who asked me: “Do you know what happens when you assume?”

Leave a Reply

Your email address will not be published. Required fields are marked *