The dot research method is very valuable. First and foremost the process is fun and engaging for both the volunteers and our customers. Our survey results are currently posted in the market every week in front of our volunteer booth. Each week people stop and read the results, then share how we could use this information to build a stronger market in our neighborhood. The process is so simple that we are continuing to do the dot survey in the market this year.
Suzanne Briggs, Hollywood Farmers’ Market
Farmers' markets play a significant and expanding role in ensuring the viability of America's small farms and the vitality of its towns. Although their importance in the food distribution system is growing, they are poorly understood. At the local level, most farmers’ markets lack the resources, the information and the skill that they need to document their role in the community and to plan for the future.
This research and extension effort has three objectives:
- Provide markets with the essential information they need to improve their operations.
- Develop research methods for conducting quantitative and qualitative research in farmers' markets.
- Provide managers and board members with opportunities to critically examine each other’s markets, develop their own analytical skills and establish a mutually reinforcing network of individuals who share information and insights.
The three objectives are interrelated – we focused initially on the first objective, established the second objective as a way to achieve the first, and later added the third to make the whole effort sustainable. We have taken to calling the combination of methods a Rapid Market Assessment (RMA) and have conducted eight RMA studies in the last two years.
The Value of Information
In a world in which Walmart has real time data on all of its operations, farmers’ markets remain at the other end of the information spectrum. In most instances, the markets serve as a meeting ground of independent business and collect no information beyond the number of vendors and the stall fees paid. Yet there is much that it would be valuable for these markets to know to help their vendors serve their customers and to help the market gain public and business community support.
As a starting point we considered standard data collection approaches for filling this information gap. We concluded that on the shelf approaches did not fill our needs:
- Random mail or phone surveys are inadequately targeted and expensive;
- In-market interviews suffer from small sample size;
- Mail-back surveys handed out in the market have very low response rate.
So we set out to develop our own methods that would be quick, inexpensive, and reliable. As a final and most ambitious goal, we wanted to develop a way of collecting information that would add to rather than detract from market atmosphere.
Act 1: Quantitative Data Collection Methods
Designing an approach to collecting market attendance information simply required developing a standardized. In our approach we carefully determine all possible entry points and then minimize the manpower requirements by counting for only ten minutes at the mid point of every hour and multiplying those counts by six to get an hourly average. This approach provides a reasonable estimate for attendance and has helped markets to better understand how their markets vary both within a season and between seasons. Nothing fancy, but it has proven very helpful to provide a single standardized approach.
To address the spending and other data collection needs we developed the dot survey approach (Lev and Stephenson 1999). It consists of constructing close-ended questions (we typically ask four carefully crafted questions per market) that are displayed on posters at a central location in the market. Shoppers are invited to answer the questions “self-service” by using stick-on "dots" (round, colorful labels). It takes participants one or two minutes to answer all four questions and go on their way (although some choose to stay and discuss the research).
Before discussing the advantages of the approach and some results, it is useful to consider three concerns. First, are participants influenced by prior answers (which are all visible)? For most questions, this simply is not an issue – where do you live, how much did you spend, even how old are you are not questions for which the participant really cares how others have responded. In fact in our experience, the response rate on difficult questions like age are higher than through other survey approaches because it is clear to the respondent that anonymity is preserved. Nonetheless several strategies can be employed to reduce this “influence” concern. We generally “seed” the posters with scattered dots and later remove these seeded dots. On certain occasions, we have removed the posters at regular intervals so that participants can place their dots with less prior information.
A second major concern with the approach is that only a few close-ended questions can be asked. This is certainly the biggest drawback and should not be underestimated. But it does force the market to focus on what information is most important and ensures that all the data gets analyzed in a timely fashion. In addition because this is a low cost approach, it is feasible (and preferable) for the market to collect information on multiple occasions.
A third concern with the approach is the difficulty (but not impossibility) of conducting cross tabs on the data. Ordinarily you do not know how answers to question 1 are related to answers to question 2. It is possible to code the dots so that this information is available.
In our view the advantages of dot surveys in farmers’ markets far outweigh the disadvantages. The approach is extraordinarily simple and inexpensive. The transparency of the research process (everyone sees everything) turns out to be one of the most appreciated aspects of the approach. It makes the research interactive rather extractive. Participants often stop back later in the market to see how the responses are shaping up. In our experience this data-gathering approach really does add to rather than detract from the overall atmosphere of the market.
This is a high volume research approach. Using it we have had as many as 900 participants in 5 hours (180 per hour). When using the approach we have kept careful records and have documented that 90% of the consumers that we approach are willing to participate in the surveys. This compares very favorably with response rates for all other survey methods. The superior quality of the sample should improve the accuracy of the data collected (Salant and Dillman 1994). We also have specifically asked consumers whether “they prefer this approach or more traditional written surveys” and 94% favor the dots.
The results can be quickly tabulated. For a Saturday market we typically provide the answers back to the market board on Monday morning. Finally as mentioned above, it is quite feasible to repeat questions and thereby have greater confidence in results.
The information obtained has been eye opening and quite useful to the markets we have studied. We always let the market studied select the question – our statewide effort suffers a bit by not having the same questions always asked but the gain in relevance to the specific market more than makes up for this. Here are three examples of the information collected with comments on why the information was useful (Lev and Stephenson, unpublished data).
|Answer||Corvallis Saturday||Albany Saturday|
Many downtown business communities are lukewarm supporters of farmers’ markets. One reason is that they don’t believe that markets really attract people downtown. In conjunction with the attendance counts, this question provides some data to address on that concern.
|Market||% Doing Add. Shopping||Average Amount Spent|
|Albany (1998 average)||38%||$4.32|
|Corvallis (1998 average)||63%||$10.13|
|Hollywood District (2000)||47%||$8.98|
This second question examines whether the people attracted to the market actually do any shopping at neighboring businesses. The extent of spillover sales depends on the attractiveness of the adjoining businesses (downtown Corvallis district has more to offer than downtown Albany). In several markets, we have found that spillover sales have been as high as 80% of in-market sales.
|Answer||Corvallis Saturday||Corvallis Wednesday||Hollywood District|
|Nothing else wanted||67%||60%||55%|
|Couldn't carry more||13%||11%||17%|
|Out of money||10%||14%||14%|
|Prices too high||6%||8%||5%|
|Ran out of time||4%||7%||8%|
This question was prompted by repeated comments by some dot survey participants that prices in farmers markets are “too high”. The question assesses whether price or other factors really do constrain sales. The results are remarkably similar across markets and the overall conclusion is clear – for most consumers prices do surface as a major factor limiting their purchases.
The markets have used the information collected in a variety of ways:
- Attendance counts and market sales:
- Demonstrate to community leaders the value and popularity that residents place on markets;
- Provide potential vendors with market size information.
- Drawing power of market and spillover sales:
- Quantify for community leaders the broader economic impact of markets;
- Demonstrate to neighboring businesses the value that the market can provide to them.
- Where customers live, why they come to the market and what stops them from making additional purchases:
- Assist market boards in allocating their budget (advertising, entertainment etc.);
- Assist vendors in selecting and pricing their products.
Act Two: Rapid Market Assessments
Act one was useful and productive. We studied several markets and distributed widely both the results and the explanation of the research methods used. While we hoped that other markets would be inspired to go out and replicate the methods (and a few did) the more common result was that the other markets simply made use of these initial results with the comment that their own market was “sort of the same”. Most managers viewed the data collection as too great a hurdle.
So we refocused our attention on developing a way to better encourage the spread of the methods. One idea that surfaced was to enlist market managers as data collectors outside of their own markets. What better way to learn than by actually doing? But then we thought once we have put the time and effort in transporting the managers, there should be a better means for taking advantage of both their expertise and their eagerness to learn about market management. Bit by bit we moved towards a participatory action research (PAR) model in which the practitioners (market managers) were fully involved in the research process (Whyte 1991).
We developed a system of recruiting teams of four or five mangers or board members from markets around the state and bringing them together to study a market on a specific day. Part of what they do is to collect the quantitative information detailed above. But they also conduct a qualitative assessment of the market. This qualitative process follows a precise set of steps.
The night before the market all of the team members gather and have dinner with the market manager and market board. The manager and the board provide a bit of market history and current situation and we have them focus on a particular issue of how they define success and where they want the market to be in five years time. The dinner provides context and sets up relationships for the next day.
At the market, the team members learn firsthand how to conduct the attendance counts and the dot survey. Once they have experience how simple and effective these methods are, they are much more likely to use them in their own markets. The team members are also given a series of three worksheets (physical setting of the market, vendors and products, and market atmosphere) to complete over the course of the market. The assignment really seems to capture their interest and involvement. Each person brings a fresh set of eyes to the task and notices different things. They are encouraged to do whatever it takes to become experts on this market – follow shoppers around, talk to them, watch vendors, and explore the neighborhood. At the end of the market we get together for a debriefing of both the qualitative and quantitative research activities. During the debriefing we talk about their main observations as well and suggestions about how to improve the research process.
Within two days of the study, the quantitative results are emailed to the studied market and the RMA team members. It is very useful to be able to provide this information quickly. The process of collating the qualitative reports is more time consuming. All of the market worksheets are typed and then one of the researchers takes on the job of compressing the five or six different reports into a single short report that follows the format of the three worksheets and has separate sections for what the observers appreciated about the market and what they either had questions about or suggestions for improving. This report is also distributed by email. After the distribution of the reports, a conference call is scheduled to discuss the top three to five issues to come out of the overall process. As an example, here are the five issues that were selected for the Hood River market this past year:
- What would be the advantages and disadvantages of reorganizing the Hood River market so that all of the agricultural producers are grouped in one section, all the prepared food vendors in another section and all the craft vendors in a third section?
- What advantages would seating near the music and prepared food provide to the market. Is on-site storage one requirement for making this feasible?
- How can the market encourage improved vendor display and signage?
- What can be done to improve links with tourism? Is cross-marketing possible?
- Any further ideas for how to attract market volunteers in a small community such as Hood River?
Overall the RMA process is exhilarating and empowering. All of the involved markets gain new insights and ideas through this two-way learning process. The markets studied have received detailed quantitative and qualitative information on important issues. The RMA team members are better connected, more analytical, and more confident about conducting research in their own markets. Based on their experience of really studying another market, they go back to their own markets full of new management ideas and options.
My experience with the assessment project was wonderful; both as a team participant in Ashland and having a team visit our Wednesday market. The information and insights were wonderful. It was time well spent and enjoyable, even in the middle of a busy market season! What a good way to meet my peers. Sometimes I feel isolated (even though I am surrounded by vendors, customers and board members!). No one really understands like another market manager. It is great to work with such a sharing, caring, dedicated bunch of folks!
Dianne Stefani-Ruff, Portland Farmers' Market
- Lev, L. and G. Stephenson. Dot Posters: A Practical Alternative to Written Questionnaires and Oral Interviews." Electronic Journal of Extension Vol. 37 #5 (1999).
- Lev, L. and Garry Stephenson. Unpublished data and reports are available from the author
- Salant, P. and D. Dillman. How to Conduct Your Own Survey. New York: John Wiley & Sons. 1994.
- Whyte, W.F. (ed.). Participatory Action Research. Newbury Park: Sage. 1991.