CPGmatters: Frito-Lay Gains Competitive Edge Using Ethnographic Research
By John Karolefski
Through
a special arrangement, presented here for discussion is a summary of
a current article from the monthly e-zine, CPGmatters.
To
better understand the mindset of consumers in the store as well their
thinking beforehand and afterwards, Frito-Lay relies on deep insights
gathered by cultural anthropologists using ethnographic research.
Anthropology
is the study of people, according to Donna Romeo, group manager, consumer
strategy & insights, shopper marketing at Frito-Lay.
“It
is holistic, which means that everything is important. Also, it’s about
context,” she said. “People are always culturally and socially situated.
There is something about them that is connected with somebody else. They
carry this context with them and it impacts how we make decisions. We
could have lots of numbers, but we need to understand the ‘why.’”
Ethnography
is an inductive research method that employs in-depth interviews, video,
analysis, participant observation, re-enactments and so forth. Frito-Lay
relies on Smart Revenue of Stamford, Conn. for its ethnographic research.
Frito
Lay outlined their 360 Shopping Process that aims to understand the food
shopping cycle from a holistic perspective using ethnography at a recent
presentation at the Shopper Insights in Action conference hosted in Chicago
by the Institute for International Research (IIR). The study was done
during the economic downturn and demonstrated the impact of the recession
on shoppers.
“We
wanted to understand what people do and when they do it,” said Ms. Romeo,
who is a cultural anthropologist. “and if ‘pre-shop’ really matters.
Is there a lot going on before people got to the store?”
Ms.
Romeo said the study sample was small – only 38 in-depth interviews. “I
believe in small samples,” she said. “They can reveal tremendous amounts
of information that can generate factors for the eventual qualitative
analysis.”
There are
three phases to today’s shopping routine:
- Pre-Shop
(make list, gather coupons, research via web and word of mouth) - Shopping
Experience (provide consistency in the store; use end caps as anchors) - Post Shop
(food is easy to store in pantry and fridge; got value via deals; pleased
family).
Nowadays,
perhaps because of the recession, Mom has taken on more responsibilities
and sometimes opts to stay at home with young children, according to
Ms. Romeo.
Her
three key insights:
- The New
Frugality (Mom bears the brunt of deciding no Disneyland vacation this
summer) - Wants versus
needs (nutrition and main meals are first) - More work
and less pleasure.
According
to the executives, the old model insufficiently combined market research,
marketing, sales, merchandising, operations, and external partners and
data sources. A new integrated model is consumer- and shopper- centric,
and internally and externally collaborative. The new model aims to optimize
pre-store and in-store media, marketing, and merchandising to increase
revenue and profits.
Discussion
Questions: What is the potential as well as limitations of using ethnographic
research to understand consumers’ mindset? What’s the most efficient
way to gain insights into planned and unplanned buys at retail?
What ethnography is very good at is describing what people do, how they do it, when they do it, and so forth. It is no better than any other technique at the WHY question. You are still relying on consumer self-report, which is biased to begin with and the bias is exacerbated by close and prolonged contact with the interviewer.
This is not a new technique and has been the basis for much of social science–you start by observing and making testable hypotheses. And with 38 people in the sample, you are not going to get a lot of deep insights.
Ethnography can provide a useful look at the consumer behavior and attitude points of interest. Best to combine this qualitative research with a quantitative research link to provide statistical reliability, and take some “guess work” out of the concern about “self reporting.”
The consumer is very good, and highly accurate in their self reporting. Who knows a person better than themselves?
Any extra data available about shopper behavior is valuable in positioning a product. Frito especially could benefit because of how their products are used. Frito’s customer base is made up of every buying group and their products are purchased for varied reasons. Examining alternate shopper behaviors can help them create more specific marketing and product presentations.
We believe in the consumer. We believe she can have a lasting impact on the success or failure of products and brands. In order to get to know the multitude of “shes” out there, ethnographic research can be very useful. This first-person view needs to be part of a qualitative AND quantitative approach. All of the data inputs need to build toward a plan of action.
My early training was in psychology and thus, I’ve always been skeptical of the sample sizes used in consumer research. As a consultant, I understand scopes and budgets, so we need to make the best conclusions we can from the data available.
At ConsumerX, we are excited about the use of ethnographic research to localize offers and assortments. This plays especially well in the consumables categories, but we are seeing interest in apparel and some hardlines categories.
I support Mr. Needel’s comments about self reporting and sample size and Mr. Saunders about increasing the validity of the research by adding additional elements to it. I believe we would all agree the “why” consumers do what they do is the most difficult to determine.
The article does not elaborate on how the 38 were selected, but the assumption I gather we are supposed to make is that they are statistically representative of Frito’s customers or potential customers. If true, this would represent an extremely carefully chosen panel.
All the negative comments aside, Frito is attempting to do what all manufacturers want to do–better understand what drives the purchase decision.
38 interviews? That’s enough?
This is a first step toward a better understanding of customer motivation, but it’s a very small step. Most retailers and manufacturers admit that they have a primitive understanding of who purchases products, and why. A deeper understanding would undoubtedly reveal actionable insights which could raise sales. For instance, why don’t salty chips come in resealable bags to keep out humidity?
Not only is the number of interviews suspect here, but also where and when the interviews took place. In the US, can you imaging the differences in insights from 38 people in Kansas, or 38 people in the Bronx, NY? Now, take that global, and you can see how irrelevant the results may become with such a small sample. Sure, there are points of diminishing returns, however, one should ensure the results are scientifically accurate.
This process is not complex. We are trying to use science to prop up those who lack the art of merchandising. While studies in human behavior are useful as are many of the tools we use in category management, I tend to believe this is but one more less than useful collection of data more designed to confuse and obfuscate the retailer. The true mission is most likely more retail display space for the study sponsor. Delving into the minutia of human behavior at store level is akin to trying to write software to predict the weather.
However, I suppose it provides an income stream for the consultants which ultimately are good for the economy. I prefer the lost art of good customer interaction at the store level and process of feedback on how we can run a motivated service team.
We’re employing some ethnographic studies to gain more valuable insight into what is a pretty complicated environment with no easy answers right now.
And what I found with a very small sampling provided some excellent insights. It’s not often you can as a manufacturer or marketer get right down there with your consumer and this is one way you can.
To do it with superior results, larger samplings are in order, but that is a massive project and I am not sure it serves the best purpose in the long run. Combining this small sampling with qualitative research is a good method, but knowing exactly what you are looking for or are trying to answer is key to knowing what you need to combine it with for best results.
Many factors should ultimately decide one’s strategy anyway. At one time we should have been able to employ the use of the retail sales clerk in these studies, but there is so little valuable interaction these days you have to find another way to do it.
I once worked in a retail store as a clerk for nearly 6 months and found out more about the next generation of shopping than any market research every presented me with. So much so, I’ve submitted my results in a proposal to Amazon in their AWS Challenge and it’s currently being reviewed as part of the competition.
Getting as close to the customer as is possible today is a much needed requirement for company management and marketers to glean valuable insight into their consumer. Ethnography is one very valuable way of doing that. Otherwise, I’d recommend putting your management in at the retail level in various venues. You’d be surprised what you learn.
Data has largely carried brands to where they are today. But quant alone isn’t enough to carry brands forward. Today’s consumers increasingly live complex lifestyles that are fast-changing. It’s definitely time for brands to marry their data with consumer insights. Only by understanding the “why” behind consumer behavior will brands be able to get closer to the consumer and thus create products, messages and shopping experiences that truly resonate with their 21st century lifestyles.
Right on Kim!
GMROI–you are correct, ethnographics are very similar to what a good merchant does–the world is a bit more complex and the need for contextual relevancy drives these ethnos. The store can be a great lab–Terry Lundgren’s an old school merchant and is into this, he’s not into paying consultants to obfuscate. Why is he using dunnhumby? Give this another look (unsolicited advice.)
Ethnos are typically conducted to get at the hows and the whys. Look for the other DAA in your CRM/loyalty data or a survey screener (who (female/demo) what (UPC # XXXXX/) when, (time of intercept) where (store location) what other technique do you use to get at an open ended why question (Why? To clean my house. Why? Because I am renting my room. Why? Because we need the income. Why? Because of the economy. Why are you buying the green brand? How did you decide…etc, whys and hows. All the other stuff from ethnos, like activity analysis, stakeholder analysis, and more, are outcomes of good research. Bias is a function of the qualifications of the contracted firm–the folks at Smart Revenue are very qualified–the IIR group is a very credible organization that has built a community around the best of the best in qualitative consumer research.
But Ethnos are time consuming/expensive (and most people don’t really do ethnos; more ethno-lite) so framing the research is key, the screener is key, and interviewing 8-12 properly selected respondents is a big job. I’ve heard folks who are more towards twenty. Ethnography (full blown) includes in-depth interviews and shadowing, home visits, a 45 minute interview transcribed is a ton of content, add videos to that and your problem will be the sadness that comes when you have to file away many great insights in favor of the lowest hanging or most strategic ones.
I recall that Smart Revenue does intercepts at the shelf (for selected demographics, say a mom with a child) that are triggered by the emotional scenario’s one is trying to study (customer picks up product, looks at another, puts it down, makes a choice.)
This type of research, at the shelf, with perhaps a “think aloud,” produces amazing content if done well, every round can help refine your screener and interview questions–the process is iterative, techniques like asking why multiple times gets you closer to the truth as well.
38 Interviews is a big number if you’ve properly designed your research plan
The location and time is all part of the plan–let’s say you decide you want more of the after-school snack business for a specific suburban audience–you set up the intercept rules, and execute (lot’s of work).
The risk of using limited sample ethnographic research is an assumption that the shopping experience is more or less consistent across segments for the important variables.
Given the growth of ethnic groups in the United States as well as rapid changes in core customer behavior across segments, I would be concerned about leveraging insights from such efforts without some validation among key business segments. How do these processes change for a single person shopping for one, and perhaps for a party, vs. a homemaker who has time for research and is feeding a family of four? How about changes in the buying process for a Hispanic mom living with her mother-in-law and 6 relatives?
Remember, we are more diverse than ever and identifying and developing personas for the different critical segments is vital. Test your assumptions across the personas–they may be perfectly valid. But you NEED to confirm–otherwise you may be painting the fence with too broad a brush!
I have lead over 100 projects for most of the top retailers in the UK and a few in the US, like Target. Most are using some form of ethnography. My take on it’s validity and benefit has evolved over the past 10 years. Retail is so complex that finding unifying “whys” is tough and critically, often not actionable. As stated above, it often needs sample sizes beyond 38. Or does it?
In my view you can’t get full value from any such study without the clients (very often a brand and retailer in partnership) being the eyes and ears of the study.
This yields “influential insight” that brands and retailers act on. In fact in this way, with just 12 precision recruited customers one can add real value.
There are 3 key ingredients.
1: seeing first hand–with some quick training–clients get to empathise with why something happens and go beyond events and influences to emotional drivers of behaviour such as stress, fear and desire.
2: This then gives them a new lens to draw meaning from existing data and previous experiences as well as focus on what can influence behaviour in a positive way.
3: Finally this needs to be landed in a mission model which is shared, evocative and critically helps those outside of the immediate challenge (sales staff, shoppers or even your family) enrich, explain and understand.
My point: the methodology is not the only issue. It’s how influential it is and how those using it apply it. Most clients have data overload. By using ethnography precisely, years of data and experience can be re applied, saving time, money, and wasted effort.