Describing Package Tours With Schema.org: Too Much Pain, Too Little Gain... If Any

Since I published some of my (mostly positive at that point) thoughts about structured data embedded in web pages in general and schema.org in particular (see Tour Shopping on the Web: the... Bad and the Ugly and Chicken Thighs, SEO, Structured Data... and More), I have been asked many times by quite a few tour industry professionals whether implementing schema.org mark-up on their sites is something they should consider.

If you want a short answer, it's in the title.

Below is a longer, minimally technical, answer. I may focus on the more technical aspects in another post when and if I have the time.

First, let me explain the following:

In order to test the benefits of schema.org markup for a web site of a small independent tour operator, a real web site was created. Real tours with very detailed itineraries and programs were developed.

The site was run without schema.org markup for about two months, i.e. long enough for it to be fully indexed and start showing consistently in search results.

After that, it was marked up with microdata using schema.org vocabulary and run that way for over five months (it takes a long time for Google to discover and "digest" embedded structured data).

Below are some of the results of the test described as "non-technically" as possible.

1. Schema.org markup will not improve your Google search ranking.

It is really amusing to watch how vague Google has been about this (as an example, see the answer by Matt Cutts, head of Google's Webspam team, below).

What's even more "amusing" is how the so-called SEO experts are eager to see the glass half-full (or, rather, persuade you that it's full) and interpret Mr.Cutts' evasive answer as a "yes". Just "google" something like "schema.org seo improve ranking" and see for yourself. I totally understand that everybody tries to make a living any way he or she can, but what really bugs me is that most of these self-proclaimed experts don't even bother to study the subject of their "expertise", let alone test what they sell. Instead, they just slap together some "snippets" from the article About rich snippets and structured data on Webmaster Tools, maybe, copy-paste a couple more sentences and primitive (often buggy) markup examples from the horrible schema.org documentation site, spice it all up with very generous amount of hype (like "the secret weapon to get ahead of the competition", "you're behind if you're not using it", and the like) and then litter the web with that part-technical part-marketing babble waiting for the naive and gullible to "bite".

Ok, that was a rant.

Anyway, just as I had expected, implementing schema.org markup did not produce any SERP ranking improvement. But there was also something I, frankly, had not expected: see #2 below.

2. Schema.org markup may cause your SERP ranking to drop.

I hate to sound like a conspiracy theorist, but it looks like, once Google "figures out" that a page represents a "product" (which, according to the schema.org's definition, is "anything that is made available for sale"), its SERP ranking goes down. Of course, I cannot prove it, but here is what makes me think I am right:

Right after I marked up some pages with schema.org tags, the marked-up pages continued to show in search results just like they did before the markup was added. After a few days, the pages got re-indexed and they started appearing as rich snippets as shown in screenshot 1 below (the SERP ranking at this point was not affected).

Screenshot 1

At that time, the Webmaster Structured Data Display Tool (not to be confused with the Structured Data Testing Tool) was not reporting any structured data discovered yet (as I have said, it takes a long time, over a month, for Google to "digest" structured data). Evidently, Google's search results display engine was "mechanically" parsing the data while the "Google's brain" in charge of "understanding" the structured data was lagging behind. This lasted for a few weeks.

When the Webmaster Structured Data Display Tool started showing structured data,

Screenshot 2
the pages immediately disappeared from the first page of search results (where they had appeared consistently before) and got pushed down to at least the second search result page.

Like I said, I am not a huge fan of conspiracy theories, but it does look like the search giant is trying to force those selling "products" to buy ads by suppressing "product pages".

To demonstrate how reluctantly Google shows "products" in search results, let me show you one more example.

Below (screenshot 3) is the very top of the first search result page. In the first position is the page that gives an overview of tours and does not contain any structured data.

Screenshot 3
And yet the actual tour (with structured data showing that it is a "product") is all the way down almost at the bottom of the second search result page (see screenshot 4).

Screenshot 4
I am not complaining about the fact that out of 9.75 million pages it is in the 17th place. I am just questioning the... logic... or whatever this kind of ranking might be based on...

3. Adding schema.org markup is quite tedious.

There are a couple of CMS's (Content Management Systems) that support some flavor of semantic markup on a basic level, which may be good enough for those whose "product" fits into what schema.org (effectively, Google) perceives as legitimate Product (also see here) and can be described using its "native" properties. Just by looking at the properties one can easily see that they are meant to describe physical goods.

In order to describe such a complex thing as a package tour, one has to use properties of other types (e.g., Event). I have not seen a CMS that does it well enough (if at all) for a non-technical user to use.

There are also quite a few microdata generators available on the Web. Those are just web forms with text boxes. You fill in the text boxes, submit the form and get the html code with microdata, which then can be pasted into a web page. The ones I have seen have the same limitations as content management systems. In addition, pasting code into a web page through a CMS (almost everybody uses some kind of CMS these days) is not as simple as it sounds because any modern CMS will, most likely, try to clean up your html, which will mess up your microdata markup.

For this test, I marked up five tours by hand, and I have to say that I would definitely not want to do something that tedious on a regular basis.

Since at the beginning of this post I promised not to get technical on you, I won't (at least, not in this post). There are quite a few issues with schema.org that are more technical in nature, and, in order to explain them clearly, I need more time, which I don't have right now.

In conclusion of this intentionally non-technical post, let me point out that, although I experimented with marking up package tours, all of the above (maybe, to a degree, with the exception of #3) most likely applies to any type of product (broadly speaking, i.e. "anything that is made available for sale"). As to #3, with some very simple products, the markup process may be less painful, but, still, the question remains: "Is it worth it?"


P.S.
As you may have noticed, all of the above is about Google. The reason I ignored the other three major search engines (Yahoo!, Bing, and Yandex) that had joined schema.org (at least, nominally) is that none of them appears to even display rich snippets. Whether they "understand" schema.org markup and use it internally to improve the quality of searches (or to "bury" pages that contain "products" the way Google seems to be doing it - see #2 above) I do not know.