Ok folks it's officially time to party. Naysayer’s’ that have been claiming this was just some PR around the Macromedia Flash Search Engine SDK and the Googlebot wasn’t indexing Flash can be silenced because we now have undeniable proof that Googlebot is really crawling links within Flash movies as quickly as regular text links in html pages.

When we heard Googlebot was indexing Flash but most of the people in the SEO community said it wasn’t, my SEO Director asked me to run some tests to find out for sure.

From our tests we now have proof that Googlebot is more than capable of following URL's contained in a Flash movie, without any modification to the movies themselves. This is food for thought for those in the community that were skeptical of whether this was the case despite the recent press releases from Adobe and Google which stated it was. In addition it seems that many SEO's weren't clear from the press releases whether the Flash authoring environment itself would be changed or the Flash player technology would play the crucial role in ensuring links would be followed.

Looking at the statement below and our test results, we can see that Googlebot itself has been effectively ‘upgraded’ to include parsing using Adobe's flash player technology. No clunky SDK required guys :-).

"Recently, we've improved the performance of this Flash indexing algorithm by integrating Adobe's Flash Player technology."
Ron Adler and Janis Stipins, Google Software Engineers

To fully clear this up however we needed to establish the following;

1. Was Google actually following links in Flash movies?
2. Would Google follow any links contained in a flash movie even if they were on a hidden layer?
3. Would we be able to use a relatively old version of Flash (Flash MX) to produce the desired results?
4. What would be the best way to embed the Flash movies so that were Google friendly?

We ran a control test which involved adding a normal text link on a PR5 domain that linked to our PHP test page. Our PHP test page fully supports the recommended guidelines set out by Google on how to verify Googlebot in order to avoid false positives, it is also set up so that we can pass variables to it in order to quickly create new tests. The script emails us immediately when Googlebot crawls it. The control link served the purpose of establishing when Googlebot had crawled the test page and that it was working.

Next we embedded some Flash movies in the footer of the website that linked to the same script using URLs with new querystring variables identifying which link had been followed from which Flash movie.

The two very simple Flash movies were created using Flash MX. Both contained a standard button and a GetURL action in layer within the movie. The button had the following action script added:

on (release) {
getURL("http://www.example.com/flash-test/scs.php?pageid=5");
}

We then added the following to keyframe 10:

getURL("http://www.example.com/flash-test/scs.php?pageid=6");

We decided to stick to absolute URLs that begin with http://www.domain.com/ initially in order to keep the test relatively simple and just prove Google could access and crawl the links. We’ll deal with relative and fully qualified links tomorrow.

The first Flash movie was embedded using standard object and embed tags:

<object classid="clsid:D27CDB6E-AE6D-11cf-96B8-444553540000" codebase="http://download.macromedia.com/pub/shockwave/cabs/flash/swflash.cab#version=6,0,29,0" width="200" height="100">
<param name="movie" value="swf/test.swf">
<param name="quality" value="high">
<embed src="swf/test.swf" quality="high" pluginspage="http://www.macromedia.com/go/getflashplayer" type="application/x-shockwave-flash" width="200" height="100"></embed></object>

A second movie was also uploaded to the test site using the standards compliant SWFObject method which uses only the object tag with JavaScript to embed flash movies. Using both of these methods would help us establish how much of an issue the depreciated embed tag would be for Googlebot, and also that both current standards compliance and search engine friendliness could be achieved.

Once this part was out of the way it was just a case of waiting for Googlebot to crawl the website. As this was a PR 5 domain we knew that it would occur on the same day. In fact Googlebot followed the standard text link at 16.18pm the same day and proceeded to follow all the URLs contained within both Flash movies between 17.29 and 18.28. This was conclusive proof that Googlebot could follow URLs in Flash movies and more interestingly URLS contained within action script code that was hidden.

Test results

1. Text link crawled: (Googlebot IP: 66.249.73.27, Time: 16.18)
2. Standard embed flash link crawled: (Googlebot IP: 66.249.73.27, Time: 17.29)
3. Standard embed hidden flash link crawled: (Googlebot IP: 66.249.73.27, Time: 17.29)
4. SWFObject flash link crawled: (Googlebot IP: 66.249.73.27, Time: 18.28)
5. SWFObject hidden flash link crawled: (Googlebot IP: 66.249.73.27, Time: 18.28)

Congratulations go out to Ron and Janis from Google for achieving this crawling milestone and also the creators of SWFObject for an excellent script which is now both a standards compliant and search engine friendly way of embedding Flash movies.

Flash Bombing? - We might as well point out the obvious as what this now opens up is a whole bag of spam possibilities and I wonder whether Google has fully considered this? Matt I know you wanted some suggestions about what to deal with next from a spam perspective, we’d suggest you gen up on Flash immediately. Before someone else does it we’d like to coin the term ‘flash bombing’. Theoretically someone could now effectively hide hundreds of links in flash movies that would be completely invisible to users and a nightmare for Googles eval and anti-spam team to verify. It’s like BM all over again ;-). Could Google fully check this? How might they approach it? Would every website using flash links need manual review? It will be interesting to see how webmasters and SEO's will respond to this.

In part two of our experiment tomorrow we will investigate how Google handles relative and fully qualified links within Flash movies and whether Google treats SWF objects as a separate entity for PageRank distribution or associates the movie with the page it is embedded within, something we are sure the engineers will have deliberated over.

Stay tuned for more details tomorrow.