It's an SEO's duty to try to utilise and leverage as many opportunities as possible for clients and employers in order to drive relevant traffic to their websites. One technique that I sometimes feel is overlooked — or at least not given the attention it deserves — is transcription, i.e. turning audio or other media into text.

I was inspired to write about transcription for SEO (and more) after talking to a client at one of my previous agency roles. A few staff members at the top of the company are well known in their industry, and we wanted to leverage their popularity and standing by encouraging them to guest blog. For one of them (who's practically a celebrity in his industry sector!), we were told this:

Client: "Well, he doesn't want to write content on a regular basis. You see, he has enough on his plate as it is with his popular, weekly, hour-long podcast."

Then the light-bulb moment happened...

Me: "Do you transcribe the podcasts into text and publish them on the site along with the audio?"

Client: "No. Why?"

Why?! Oh, my sweet, naïve client...
(I didn't actually say that in reply to the client! That'd be silly.)

Ahem... Where was I?

Quite fittingly, my first instance of seeing regularly transcribed content was on this very site: Moz's Whiteboard Friday videos are all transcribed on a weekly basis (or at least they have been every week for the last few years).

For that reason, it only seemed right to talk about transcription in the form of a YouMoz submission!

I think there are benefits beyond SEO, as it also touches upon user experience (UX), and if you sit down and really think about it, there are a lot of different things you can transcribe, which is why I've also provided a list of ideas towards the end of this post.

The benefits of transcription for SEO

The main benefit of transcribing audio for SEO? Search engines cannot 'read' audio media. Yet. Properly.

Yes, you can add text to an image to help search engines deduce its content and purpose (in the form of the title and alt attributes), but that's not necessarily the case with things like videos. Embed a YouTube video, look at the code and see for yourself — it isn't full of the video's text, ready to be crawled by a search engine spider.

And while search engines are getting wiser and more Skynet-esque by the day, they're still a long way off from effectively turning audio into words. I can't find the exact tweet right now, but someone recently tweeted @mattcutts asking if the Webmaster Videos were transcribed. He replied saying that they were automatically transcribed on YouTube, accessed via the "Transcript" button.

I checked a few of Matt's videos and they weren't too bad, but what about when the audio isn't crystal clear and/or the speaker has a bit of an accent? I checked a video I made using my laptop's webcam and inbuilt microphone, spoken with my unusual accent (which I've been told sounds Welsh, Cockney, and accentless all at the same time), and found that the line:

"...in this video I'm gonna talk you through how to implement rel author..."

had been transcribed into:

"...video onions will keep you have to impeachment gravel for..."

Nailed it. (And I honestly thought I spoke quite clearly in that video!)

So I think it's safe to say for now that transcription through a more — how to put this — "traditional" method (i.e. through transcription service providers) is still essential at this stage.

The major benefit of transcription for SEO? Hitting the long tail. What if a video or podcast covers a topic that's not talked about in a blog post or other supportive text? Or, what if people are searching for a spoken quotation, as opposed to a written text quotation? Without transcription, they'll miss it. With transcription, they won't.

When I created the previously linked-to video about impeaching — er, I mean implementing rel="author", I embedded it in a post on my own blog along with the transcription, potentially driving more people to my blog from organic search — especially those searching for something relevant to the video and/or the event at which I spoke.

Another good example: the Q&A at an event after a speaker has given their presentation. The speaker may share their slides and speaker notes, but Q&A is obviously quite impromptu and on-the-spot in nature. If a video has caught it, and that video has been transcribed, then people looking for the answer to one of the questions that was asked will be able to find it.

The benefits of transcription for UX

I also think that there are more benefits to transcription than just improving long tail SEO. It can vastly improve usability and UX, too.

There have been numerous times when I've wanted to watch a Moz Whiteboard Friday, but I've been in a public place and not had any headphones. The next best thing? I could read the transcript. In fact, some people I've spoken to prefer to read a transcript than watch or listen to something. Each to their own, I guess, but at least by providing both you're giving your users the choice.

Additionally, when I revisit the Whiteboard Friday at a later stage and want to double-check something that Rand or whoever has said, I can use my browser's "find" function, type in the relevant word(s) and find it right away. So it's good for quick checks and references as well — much quicker than trying to find the exact moment in a 5-10 minute video when something was mentioned.

How to do it (and is it really worth it?)

I'm sure that there are plenty of transcription service providers out there. Wanting to try it out myself, I went for Moz's provider: SpeechPad. It seemed pretty reasonable and I had no major problems with it. I had to tidy up a bit of the text (e.g. Gafyn's name — which is the Welsh variation of Gavin — was spelt the non-Welsh way, some Twitter handles had been missed, etc.), but it was about 95%+ correct. All in all, $5 to transcribe my 5-minute rel="author" YouTube video? Bargain.

I know what you might be thinking: Is it worth it if a) you produce (or have previously produced) lots of media, or b) your media is quite long, e.g. an hour-long podcast or an event?

Well, put it this way. I paid $5 for a 5-minute video to be turned into text, which was 645 words long. It's unique text, and apart from a bit of a proofread and tidy-up afterwards, it was good to go. I know people who pay 10 times that amount (if not more) for 600 words of unique content. When you look at it that way, it's pretty reasonable. An hour-long transcription is likely to be essay-sized — in the 1,000s of words — which should hit the long tail like crazy.

Transcribe ALL the things! A list of things to consider transcribing

The list of things that you can transcribe is pretty much endless, so I wanted to put a shorter list together to spark ideas and make you think of what your business or your clients might have produced already that is transcribeable (and if that's not a word, I'm totally coining it):

Events

• Presentations, panels, keynotes
• Q&A
• Vox pops between sessions
• PR stunts (if they're filmed)

TV & Radio

• Full TV/radio shows
• Appearances on TV/radio shows (e.g. if your client only appears on a five-minute segment)
• Adverts

Podcasts

• Full podcasts
• Appearances on podcasts

Music

• Lyrics (especially if it's an unsigned band — they might not yet have their lyrics plastered on every lyrics website ever)
• Live shows (especially if there's banter between songs and/or alternative lyrics)

Other

• Interviews
• Whiteboard videos (obviously!)
• Corporate/promotional videos
• Testimonial videos (as in testimonials from clients/customers)
• Webinars
• Google+ Hangouts (I'm thinking #maximpact...)
• Videos with commentary/voice-overs
• Documentaries
• Pretty much everything/anything that has (or could have) audio!

Have I missed anything obvious? I'm sure I have! If you think of anything that I might've missed, leave a comment below!

Now if you don't mind, I'm off to video my gravel and impeach some onions... or was that the other way around?