Close this search box.

Could Mechanical Turk be the best content marketing research tool possible?

Last Updated on August 21, 2019 by Paul Metcalfe

What if I told you that for eleven measly dollars you could find out what your target audience wants to read and where they look for articles about that topic? Before I sat down to write this blog post, I spent that much on a club sandwich. And where is that sandwich now? It’s gone, but $11 worth of content marketing research could lead to a long lasting return on investment.

I’m a Mechanical Turk evangelist. I love it for research and for paying a ton of people a dime each to do little bits of work for me. Mechanical Turk, or mTurk for short, is a digital workforce organized and run by Amazon. Most people don’t know about it, and of those who do know of it, most don’t know how to really use it. Unfortunately, it’s not the easiest system to use in the world, but it’s definitely not impossible at all.

Who is on mTurk?

Well, it’s a huge workforce composed of people of various work background, demographics, and hobbies. Some are currently employed or underemployed, and some are in between jobs, or stay at home moms trying to make extra money as they can, et cetera. Mechanical Turk workers, or Turkers, are on average better educated than the general American population. You’d be surprised who you can find on there. I’ve recently run surveys targeting online daters, student pilots, graphic designers, and even startup founders, and had no issue finding people to quickly take a survey.

What can you learn from them?

I founded Vacord Screen Printing, but I also have a couple of startups, including an idea that would??train student pilots for the written FAA exams they take in order to get licensed. Not only was I able to use surveys to determine that 75+% of the survey takers want a system like I designed, but they’ll pay handsomely for it. Mechanical Turk is awesome for idea validation like that (see my article on validating an idea with mturk), but for content marketing reasons, I also asked the student pilots what confused them most about studying for these exams. I can take their responses and shape my content marketing.

Ask your target market what they want to learn and:

-Build an e-book on the topics that interests them and distribute it

-Write blog posts about the topics and then acquire inbound links to the blog

-Say to larger blogs that “80% of your audience wants to learn more about XYZ when surveyed” and get a guest blogging spot to get more traffic to your own site or blog

Perhaps even more brilliant than that, you could go a step further and ask your survey takers not only what they want to read, but where they would go to read it. I did that with the graphic designers, and I found out what they would want to learn about screen printing, and what sites they would visit to try to find these articles. So now I know what they want to read and where to put it so they can find it!

Ask your target market where they go to learn about your topic and:

-Get lists of sites to try to put your articles on

-Find out what publications they really read and try to get articles or mentions there

-Learn of specific influencers that they respect, and try to coordinate efforts with them

How do you get the surveys on mTurk?

The nitty gritty. I will try to give a thorough but brief overview:

  1. First, you have to sign up as a “requester” on mTurk.
  2. Fund your account with some money
  3. Design your “HIT” or Human Intelligence Task. You can use the description space to limit who takes your survey (“Moms thinking about returning to the workforce wanted for 7 question survey”)
  4. Determine a price. I pay $0.50 for people to take surveys. (Amazon adds a 10% surcharge, so 20 surveys cost me $11.) I think that’s more than enough. You could pay less, but your HITs will get filled slower, and you might get less quality responses. Also, don’t forget that the Turkers themselves are a very organized workforce, and they’ll review you on their own sites and say whether you’re good to work for or not!
  5. Limit by country. I limit to the United States, but adding Canada too makes sense, and a Turker actually emailed me a couple days asking me to start adding Canada too. But limiting to the USA/Canada will keep people in foreign countries from taking the survey.
  6. Don’t require people be “Masters” to take your HITs. It costs extra and is not at all necessary. Amazon will try to get you to turn it back on. Don’t.
  7. Design your HIT. This is the tricky part, because it’s designed in HTML, and most people aren’t experienced with that. I posted my code I use below.
  8. Make sure you link your HIT to a survey somewhere else. I use SurveyMonkey because I can have a nine question survey for free (see tips section below to see why it’s only nine questions).
  9. Publish and sit back and wait.

Here is the code that I generally use to create the HIT:

Please follow the link for the survey

Please feel free to be as long winded and verbose as you want to be in your answers in the survey when there is a comment box! The more we learn, the better we can improve. This is a survey about being a student pilot or pilot and taking FAA written exams. The survey consists of just seven questions. Only US citizens should take this survey, as the service will be based on the FAA tests. You need to have be involved in aviation to be able to accurately take this survey. Take the survey here and then enter the code that you receive after finishing the survey into the text box below. Thank you!

Tips for effective surveying

  • How not to get ripped off – I request that the turkers put in a special code after finishing the survey. That way I can tell if they actually finished the survey. Most turkers take it very very seriously, and want to keep their approval rating high, but I have run into a couple jokers (less than .1% of workers, probably). Make your survey two pages, and make the first page all your questions, and the second page a thank you page that has a code, such as “Thank you for taking the survey! Please go put the code word BANANA in the HIT to show you completed it.” You can do a 10 item survey for free on survey monkey, so this lets you have 9 questions and then the 10th item be a second page with the special code.
  • Keywords are super important to having your survey be able to be found. When I was prototyping my startup Anterose Online Dating Profile Critique, it would take days to get 10 people to review a profile and give feedback on it. Once I added proper keywords, it started taking as little as 45 minutes to fill all the reviews. Huge huge difference there. Most popular keywords that I try to always include would be “fun, easy, english, survey, quick, online, feedback” and so on, and then I add things more relevant to the survey like “graphic design, graphic designer, design, art” and so on.
  • Encourage people to be verbose in their responses. That makes a difference. Also keep all the answers open ended by having a big comment box for a response, not a rating.

I personally like written surveys, but phone surveys can be even better for in-depth research. Customer Discovery Ninja makes running phone surveys via mTurk a breeze, and includes tools that make the interviewing process easier too. Their system automatically records the call, so you don’t have to worry about taking notes and can just let the conversation unfold.

So that’s my two cents on how you get a ton of inspiration for your content marketing campaign without spending that many cents.

Have any experience with this sort of research? Have questions? Leave a comment below!

Did you get something out of this article? The best way to say “thanks” is to share the article, so please post on your preferred social media outlet.

More from the blog...

Last Updated on July 8, 2024 by Adriaan Starting a new business is an exciting journey...

Last Updated on June 28, 2024 by Adriaan The quality of customer experience can make or...

Last Updated on May 10, 2024 by Nick Startup failures can be devastating. Founders face the...