The reason that this month’s newsletter is somewhat late is that I have been analyzing and interpreting market research data for the last three weeks.

As some of you are aware, on our main website, we were offering free samples of Qeva Joint Mobility, a natural product that helps dogs with joint problems and arthritis. In order to get a free sample, dog owners had to complete a pre-treatment survey on their dogs as well as a post-treatment questionnaire. We received over 3,500 “pre” surveys and nearly 300 “post” surveys.

In this article, I will share with you some of the things we learned doing this interesting market research project.

I used the requests for free samples as a measure of interest of dog owners in dog health products. Most of the people requesting samples had one or more dogs that were suffering from arthritis and joint problems. I say most, because some dog owners were asking for the product for preventative reasons. Therefore, the data provided by these “requesters” constitutes a pretty good profile of the potential consumers of dog health products.

So what geographical areas show the greatest potential for dog health products? Of all the requests for free samples, 92% came from the United States, 5% from Canada and 3% from a dozen other countries. In the United States, the top states were Florida, California, Texas, Pennsylvania, New York, Ohio, Michigan, North Carolina and Washington. We did receive requests from all 50 states and 10 Canadian provinces. If you are selling a dog health product, you know now where you should focus your marketing efforts.

I wanted to know whether demographic data could help us predict areas of high interest. I took certain state demographic data from the US 2000 census (a tremendous resource, and it is available free at http://www.census.gov/main/www/cen2000.html) and “correlated” them with the number of requests we received from each state. The best predictors of interest were the number of households (r=.919), number of families (r=.916), number of single dwelling homes (r=.909), number of employed (r=.899) and state population (r=.895). A Pearson correlation coefficient or “r” of 1.000 means a perfect relationship. Thus, all of these correlations are very high. What does it mean? If you are looking for a place to sell dog health products, look for a region with the highest number of households.

Interestingly enough, there was NO relationship between requests and median household or family income for that state. Money is not a predictor. It appears that dog health products are a general consumer item, rather than a niche product.

Using the same demographic data, we decided to see whether any regions of the country showed greater interest in our products relative to the population and number of households. We found that the Southeast and Mid-Atlantic states were over-represented. These are the regions that I would focus on first in my marketing initiatives. The Northeast and Western states showed the least interest based on requests relative to population.

Where did people find out about our free samples? Most said from a friend or relative. The rest were from the Internet including websites, search engines, just surfing, e-mail, newsletters and so on. Our requests really skyrocketed when our offer was posted on several “freebie” sites or websites that compile information on free offers. Two that were frequently mentioned were Freaky Freddies at http://www.freakyfreddies.com/pet.htm and ArcaMax at http://www.arcamax.com . If you have free products that you want people on the Internet to know about, you should check out these sites.

What breeds of dogs were most frequently suffering from arthritis? Over 30% of the dogs in our free sample requests were of a “mixed” breed. The next most common breeds were Labrador (10%), followed by German Shepherd, Rottweiler, Greyhound, Golden Retriever, Cocker Spaniel, Chihuahua, Dachshund, Poodle, Chow Chow, and numerous other breeds. We suspect that this distribution represents the most common breeds of dogs owned by Americans, rather than the propensity of any one breed to have joint problems.

The really fun part about market research is that it often shows just how wrong my assumptions are. In promoting Qeva, I was focusing most of my energy on dog and kennel clubs. I figured this was the best way to get the word out to dog owners. Once I had a little more money, I was planning to buy some ads in dog magazines. Bad idea!

We asked the dog owners in our survey to list their main sources of information about dog health products. Topping the list, at 68%, was their veterinarian. Second, at 60%, was the Internet, followed by other dog owners at 30%. My first choice – dog clubs – came in at a lousy 2%. Dog magazines came in at 17%. Salespeople at pet stores were the source of information for 11% of the respondents, while salespeople at stores like Wal-Mart were at 4%.

Providing information to vets and having a good, informative website are good investments if you are selling dog health products. Contacting dog clubs or advertising in their publications is a waste of money, as is advertising in dog magazines. Also, don’t count on salespeople, especially at the mass retailers, to have much knowledge about your product.

Our other major plan was to try and get veterinary clinics to carry and sell our product. Wrong choice again! Although our respondents said that their vet was the most important source of information about dog health products, vet clinics are the least preferred place to purchase products. Only 4% of our respondents said they prefer to purchase their dog health products at their vet clinic.

For preferred purchase locations, mass retailers such as Wal-Mart are the top of the list (37%). Next at 29% is the local pet store, followed by on-line purchases at 21%. Since about half of American households have access to the Internet, this translates to about 10% of the general American population that are willing to purchase health products on-line. The number of people shopping online is increasing all the time. But for now, if you really want to be a serious player in this marketplace, your product needs to be in pet stores and mass retailers. However, we all know how tough getting into these distribution channels can be.

Issues with this research approach

Using a combination of free offers and online surveys in this project presented some interesting challenges and raised some important issues. If you are planning to use this approach, here are some lessons we learned.

1. Offering free samples with a requirement to provide data does appear to be a good way to do market research and provide exposure to your product. However, it does take some time, effort and luck for your offer to become known by Internet users. High placement in the search engines and use of the various “freebie” sites can help gain this exposure.

2. There were abuses of the free offer. People were requesting samples for puppies and for perfectly healthy dogs. Even though we said one offer per family, people were requesting samples for up to five of their dogs. A few were submitting multiple requests. Also, we suspect that any samples sent to Russia and Ukraine would simply be sold on the black market.

Some people did not provide valid e-mail addresses. This is understandable, as they do not want to be spammed. However, it makes it impossible to contact these people to remind them to fill out the post-treatment questionnaire. (We had some 450 e-mails bounce back out of the 3,700 that were sent out as reminders).

To prevent these abuses, we suggest that researchers take the following actions. First, make the terms and conditions upon which the free sample is being provided absolutely clear. Tell people that they will only get the free sample if they meet certain criteria (e.g., their dog has been diagnosed as having arthritis by a vet). Also use, and tell people that you will be using, a confirmation e-mail. Upon receiving a request, an e-mail is sent to the person to confirm that, in fact, they want the offer. If the e-mail bounces, then the request is automatically discarded. This also prevents people from ordering on someone else’s behalf.

The requests for free offers should be carefully screened. Those that are unsuitable should receive an e-mail stating the reasons. Careful track should be kept of the people that are sent free offers. A follow-up should be made after an appropriate time to obtain the second round of data.

3. Handling the volume of data generated in this study was a real headache. It is absolutely critical that all the survey data be captured electronically. However, it is more important to structure the responses to the survey questions in order to minimize the work required prior to analysis. We made the mistake of using too many open-ended questions (e.g., breed, state, country, etc.). In order to tabulate the data, we needed consistency. Dozens of hours were spent just getting the state and province data consistent so that it could be counted. We strongly recommend the use of as many radio buttons, check boxes or drop-down menus as possible to capture responses to questions. Also, try and keep the responses numerical – it makes it easier to analyze and allows for advanced statistical tests. The more recent versions of SPSS are pretty good in importing data in various format, but their “cleansing” tools are somewhat limited.

The other challenge was the matching of pre-post data. We did this manually by finding and adding the pre-treatment data to the post-treatment surveys. If we had received considerably more survey returns, this would have been a very time-consuming task.

4. Another issue of concern is the validity of the data and how representative is it of our desired target audience. Unfortunately, there is no way to be completely sure that your respondents are truly representative of the larger population. In our analysis we took steps to constantly compare the demographics to try and detect any biases. Other than the over-representation of Greyhounds in our first pilot, we are quite confident that our findings can be generalized to the larger population of American dogs and their owners that need health care products. Also, the results to date from our sales confirm that our conclusions and predictions were correct.

5. Are free samples a good way to market dog health products? Many companies use this strategy to introduce consumers to their product. Based on our online sales so far, we have not recovered the costs associated with this program. However, the program did create an awareness of our product, provided much useful data and facilitated the establishment of other distribution channels such as retail and pet stores. At this point in time, it is difficult to calculate the cost-benefits of this approach.

In each of your businesses, collect and use as much market research as possible. It won’t necessarily give you all the right answers, but it can certainly help you make better decisions and reduce your risks.