« Recent Announcement of Canadian Cell Phone Spectrum Set Aside for Newcomers is Political Bullshit | Home | Trust Network Parties and How They Help Your Consulting Practice »

ProTip: How to Get World Class Intelligence for $3,000 or Less

By crooky | December 5, 2007

Big market research firms like Ipsos Reid will charge tens of thousands of dollars for a market research project. You can achieve the same or better results with some affordable tools, the right marketing approach and a little know-how.

Step 1: Design a clever survey with little or no qualitative questions and a clear idea of how you want to analyze the data afterwards. A survey is a great tool to summarize the opinions and experiences of a large group of people. The bigger the group, the more reliably the survey can predict group behavior. For example - if you’re looking for a survey that predicts whether or not the general public in Canada would buy your new gadget, you need general national consumer behavior - no need to focus on specific communities. Any regional differences in behavior will shake out of your data if you get a good response rate. Make sure you keep your survey under 30 questions.

Step 2: Post your cleverly designed survey on a low-cost internet-based survey tool (watch out for the Privacy Act - if you’re collecting names and e-mails from Canadian citizens, you have to use a survey tool whose servers are in Canada).

Step 3: Buy a prize to give away in a draw of survey respondents (electronics between $150 and $300 are always a good incentive).

Step 4: Promote your survey on sites like ContestCanada.com and web-sites associated with the subject area you’re studying.

Step 5: Run the survey for 6-8 weeks and voila! If it’s a well-designed survey and an interesting topic, you should get more than enough respondents to generate statistically accurate data. (see the link at the bottom of this article for more information on what constitutes statistically accurate survey results).

Step 6: Run some sophisticated analysis on your data (eg. Regressions and Cross-Tabulations) to get at the real nuggets of intelligence that your competitors would give their BlackBerry for.

Step 7: Distribute your prize and pat yourself on the back.

“How can it be that easy?” - I can hear your incredulous doubt already. The magic is in the survey design (a topic we’ll cover in later weeks) and making sure that you incentivize your survey with a prize of some kind.

If you haven’t heard about “selection bias”, it’s something that people more familiar with market research will be muttering right now. The theory of selection bias proposes that if something motivates someone to participate in your survey in any way, it introduces a bias into your data that may mean your results are not reliable.

I’m going to piss off many research scholars now by suggesting that selection bias is a load of crap when it comes to web-based surveys with a prize. I’ve heard the following arguments:

  • People will only fill out the survey because they want the prize. My answer: So what? As long as they don’t lie on the survey (and why would they bother?) you’re still going to get the responses you need.
  • Not everyone has internet access. My answer: that’s increasingly a lame argument. Internet penetration in Canada is now over 85%. That’s rapidly approaching telephone penetration into households.
  • You’re going to bias it in favour of “internet people”. My answer: There is no “internet demographic” - everyone uses the ‘net. I’ve run prior web surveys where the demographics from 700+ respondents matched very close to Statistics Canada demographics right up to people in their late 50s. Unless you’re doing research on policies or products targeting senior citizens - you don’t need to worry about this. Besides, I always get a healthy smattering of respondents well into their late 70s on every survey I run.

    Futhermore, a lot of people don’t participate in telephone surveys anymore because of the inconvenience factor - you can’t complete a phone survey when it’s convenient for you. I would suggest that most people would be more willing to fill out a web survey than a telephone survey.

    • By way of an example, I conducted a survey in this way for a client who is developing a new platform for pushing multimedia files to cellphones at no cost to the user using Bluetooth. We had a 30 question survey that I ran online, offering up Best Buy gift cards as an incentive. The response rate was well over 700 people. One of the advisors to this client was a former executive from Ipsos Reid and when I showed him the results, he was blown away. For $3,000 (including my time), I had delivered market intelligence that was at least as good as the data Ipsos Reid could generate for a lot more money.

    So, I hope I’ve demonstrated that world class market intelligence is within the reach of even the smallest company. If you’re interested in more information on these kinds of surveys or are interested in designing one for your organization, contact me.

    If you want to know more about surveys, statistics from surveys and response rates, check out my Surveys 101 page.

    *********************

    Aaron “Crooky” Cruikshank is the Principal and Founder of Friuch Consulting. He has written professionally about science and technology for ten years.

    Share/Save/Bookmark

    Topics: Research Methodologies |

    Comments