Documentation/Rich Snippets

Study: Site Owners Are Too Impatient When It Comes to Google Rich Snippets

Abstract: As part of our Semantic SEO research, we analyzed the behavior of Web shop operators when it comes to using structured data markup. e.g. GoodRelations or schema.org, for getting Google Rich Snippets in the Search Engine Results Pages (SERPs). We can see very clearly that the number one reason for failure is that site owners are too impatient and remove the markup way too early, likely because they do not see the intended effect in Google. A clear majority of the shops we analyzed shows Rich Snippets if the markup has been present for at least 30 days.

Overview

Google Rich Snippets

Rich Snippets are an important feature of the Google search engine. A Rich Snippet is a set of additonal, likely relevant details about a page shown directly in the Search Engine Results Pages (SERPs). Typical examples are Rich Snippets for products, which can include

  • rating information (stars and number of reviews),
  • price and currency, and
  • availability.

Product picture data is being crawled by Google but not currently shown, likely for performance reasons.

Extension Modules for Shop Software

There are many ways of providing the data for the search engines, including the usage of schema.org and GoodRelations data markup in your article detail pages. The fastest and simplest way, by far, of adding Rich Snippets to a shop site is using one of the free GoodRelations modules for shop software packages, as listed here:

http://wiki.goodrelations-vocabulary.org/Shop_extensions

Goal of this Study

In this study, we want to find out

  1. how site owners use the free extension modules,
  2. if they are successful in getting Google Rich Snippets, and
  3. the pitfalls and causes of not getting Google Rich Snippets.

Method and Data

We took a list of 1181 shop sites running the Prestahop shop software in which we had at least once spotted the installation of the respective GoodRelations extension. The source of this information could have been either our regular crawls of Web data or the ping information sent to http://gr-notify.appspot.com/.

From the list of sites, we drew a small, random sample of n = 20 using the random number generator at random.org.

Then, we measured the following site features:

  1. Is the shop site operational as of today? Ideally, an HTTP request to the shop main page should return a 200 status code.
  2. Does the shop, as of today, contain GoodRelations or other data markup? For this, we took the time from the first day of finding GoodRelations markup to the date of the analysis as an estimate, i.e., we assume that a shop has had the markup without interruption.
  3. Does the data markup validate with the Google Rich Snippet testing tool?
  4. For how long has the GoodRelations markup been accessible to search engines?

Results

  1. Of the 20 shops in the sample, 90 % (n=18) were publicly available at the date of the analysis. 10 % (n=2) were off-line. One shop (5 %) was available to browsers but yielded an HTTP error (status code 406) for typical HTTP requests and was thus also not accessible to the Google crawler. So 85 % of the shops (n=17) in the sample were correctly deployed and basically accessible to search engines and browsers.
  1. Of the available shops in the sample (n=17), only 29 % (n=5) had GoodRelations or schema.org markup still included in their product detail page markup. This is insofar surprising as initially, all shops in our population had at least tried rich markup. So for 65 % of the shops in our sample, the rich data markup has been removed in the meantime.
  1. Of the three shops that had the markup available to the search engines for at least 30 days (which is estimated typical delay for Google to turn on Rich Snippets), two reliably show Rich Snippets in the SERPs. This means that 66 % of the shops in our sample that were patient enough to wait for the effect finally got Rich Snippets in Google.

The following table shows the estimated duration for which the shops with rich markup in the sample had the markup available to search engines:


Table 1. Duration of availability of rich markup to search engines
ID Top-level Domain Duration of availability of RDFa markup (days)  Rich Snippets shown in Google?
1 .pl 115 yes
2 .co.za 114 no
3 .cz 33 yes
4 .com 20 no
5 .net 8 no

Discussion

The majority of shop sites in the sample (66 %) that

  1. added GoodRelations markup in RDFa and
  2. kept the amount for a significant amount of time (>30 days)

finally got Rich Snippets in Google.

Two shops that had added the markup just 12 and 20 days before our analysis do not yet show Rich Snippets.

As for the one shop from .co.za that does not show Rich Snippets in Google despite correct markup in RDFa and availability for more than three months we can only speculate, but we assume that one of the many factors used in Google's component for checking the likelihood fo trustworthy markup may have failed in this case, e.g. due to the rather exotic top-level domain (.co.za) or the nature of the goods.

We assume that with a larger sample, the total success rate will be even a bit higher, ranging in the 70 - 80 % percent range. Note that all shop extensions use RDFa in the so-called "snippet style" (PDF with background), separating the data markup from the visible content. This is the only practical way of adding this automatically. Our study clearly shows that despite common hearsay (and some confusing automated responses by Google), RDFa in snippet style is fully accepted by Google for Rich Snippets and not considered "hidden markup". 

The main surprise is, however, that so many shop owners (70 %, n=11) removed rich data markup in RDFa by deinstalling the Prestashop extension before the SEO effect in Google could materialize. While we have no evidence to support this, we assume that many of them will have hoped for a quick success, i.e. by adding the markup and seeing the positive effect in Google immediately or within a few days time. While there may also be unknown bugs in the software that caused the removal, we assume that this is a lesser cause, since we know from hundreds of successful deployments of this very module that is works without any problems.

The important message to shop operators regarding rich e-commerce markup and Google Rich Snippets is thus: Do not give up too early! It takes 30 or more days for Google to turn on Rich Snippets for your site. A more detailed description of why Google may not show Rich Snippets for your site is here.

References

  1. GoodRelations, http://purl.org/goodrelations/
  2. Google Webmaster Tools: Rich snippets not appearing
  3. Hepp, Martin; García, Roberto; Radinger, Andreas: RDF2RDFa: Turning RDF into Snippets for Copy-and-Paste, Technical Report TR-2009-01, 2009. PDF at http://www.heppnetz.de/files/RDF2RDFa-TR.pdf.

Category:User Guide