Virginia Food Group case study
Finding growth beyond gifting season
How Virginia Food Group used measurement to identify promising markets for everyday buyers.
The case at a glance
The challenge
What we learned
The result
Gifting season brings a surge of orders to Virginia Food Group. The harder question: how could the business grow during the other nine months? To find that opportunity, the team needed to understand the customers buying for everyday occasions, what brought them to the brands, and where to find more people like them.
Who is Virginia Food Group?
Virginia Food Group brings together three brands: VA Diner, Peanut Shop and Feridies. The family-owned portfolio specializes in premium peanuts, nuts and food gifts, sold through its e-commerce sites, wholesale partners and corporate gifting programs.
“We sort of know how to win November and December. What I don't know is how to win March.”
A business built on one season
Growing beyond gifting season required understanding what held sales back during the rest of the year. The brand wanted to know whether more advertising could help, or whether the business needed to address something else first.
“Before I put a dollar behind the media, I need to know: is this even a media problem?”
Four methods, each answering a different question
The work connected an understanding of the business with a search for new customers, then tested whether advertising could generate additional sales.
Marketing accounting framework
Understand how each brand and customer segment contributes to the business.
Marketing mix modeling
Identify what drives sales across those brands and customer segments.
Geospatial analysis
Find promising markets for everyday buyers.
Incrementality testing
Measure whether advertising generates sales that would not otherwise happen.
Method: marketing accounting framework
Virginia Food Group’s three brands span different customer groups and sales channels, each with its own economics and buying patterns. The work separated the portfolio into eleven P&Ls, distinguishing new and returning customers, corporate accounts and Amazon where relevant. This made it possible to assess each part of the business separately and see differences that a portfolio-wide view could conceal.
Method: marketing mix modeling
Each of the eleven P&Ls received its own marketing mix model to understand what drove sales within that segment. The models estimated the contribution of baseline demand, category demand and paid media, giving the brand a clearer view of how those drivers varied across the portfolio.
| Brand | Organic baseline | Category demand | Paid media |
|---|---|---|---|
| VA Diner | 73% | 23% | 4% |
| Peanut Shop | 59% | 22% | 19% |
| Feridies | 36% | 38% | 26% |
Method: geospatial analysis
The next question concerned where to find more everyday buyers. Geospatial analysis examined purchasing patterns outside gifting season, then compared customer locations with demographic data to identify promising markets.
01 Aggregate
Group first-party sales by ZIP code and separate everyday purchasing from gifting-season patterns.
02 Compare
Compare those patterns with census data to identify characteristics associated with stronger sales.
03 Score
Assess ZIP codes for customer fit and compare that potential with existing sales.
Buyer profiles
Gifting buyer
Geographic patterns associated with gifting purchases: location and urban density.
Everyday buyer
Geographic patterns associated with everyday purchases: household income, home value and education.
The everyday-buyer profile highlighted opportunities within existing markets and an additional market with no recorded sales. Its demographic fit made it a candidate for testing whether advertising could generate new demand.
“Strip out the gifting weeks, rebuild on the everyday signal, and a different buyer shows up.”
Method: incrementality testing
The next step tested whether advertising could generate additional sales in the market identified through geospatial analysis. A connected TV campaign with Universal Ads compared performance in the advertised market with a comparable holdout market where the campaign did not run.
| Segment | Estimated lift | Including estimated carryover |
|---|---|---|
| New consumers | +98.3% | +109.0% |
| Returning consumers | +30.9% | +34.3% |
| All consumers | +52.9% | +58.7% |
| Amazon | +7.9% | +8.8% |
| Total, revenue-weighted | +33.1% | +36.8% |
“Geo tells me where. MMM tells me what's already working. Incrementality proves what will actually move the needle.”
A clearer path beyond gifting season
The work connected an understanding of Virginia Food Group’s existing business with a new market opportunity and a test of advertising’s impact. Together, those findings gave the brand evidence to guide its next investment in reaching everyday buyers.
Estimated 36.8% revenue-weighted lift.
“I'm not guessing anymore. I have a causal, testable path from a gifting moment to an all-season habit.”
Watch the case study discussion
Read the video transcript
Seasonally, we were a giftable company, but we wanted to go to more of an everyday consumer and get their money every day, all year round. I don't have unlimited money. We're not a billion dollar business.
I can spend as much money as I want as long as I have results behind it. So I don't just want to throw a bunch of national campaigns out there. And this is where these guys were able to help out. When I say tech stack,
there was no tech stack. There was some old Jenga tiles on the floor. The M-squared team has been great at making what they could off of the old data while we were getting the new data in there. Within that there are three different purchasing
experiences, right? There's direct-to-consumer through their website. There's corporate sales, and their sales on Amazon. It looks like there are 11 P&Ls.
That will be 11 models, and that's what we did. Feridies consumer new, the biggest opportunity: 59% driving new sales through their own website.
That really is the goldmine for them. The product is very similar, and in a lot of cases almost exactly the same. The fact that these are so different. This was eye-opening in ways that we could attack
the companies differently. Your challenge was, I know how to win in holiday. How do I win outside of holiday? And what insights can I glean from geospatial if we strip out the
gifting weeks from the model and just profile those other times of the year to identify the high propensity zip codes that are going to work hardest for us. We launched an incrementality test
with our partners at Universal Ads. Universal Ads launched the end of July, ran a six-week test and ran it in markets where you typically don't spend your marketing budget. How can I drive incremental
growth in pockets of the country who aren't aware of us? Within three or four days, I could see the difference. By the third week, I'm like excited. I'm the person on the phone with my CFO going,
aha, I told you, I told you, this is why you need to give me some money. We didn't do a mid flight read or anything, but you just knew looking at your own internal data. Maybe it was correlational.
We needed to prove it out causally, but you knew this was working. There was maybe $3,000 worth of sales that were attributed off of platform, where we saw closer to $50,000 worth of sales that are actually coming from it.
It all started with that MAF: segmenting into the right P&Ls for their specific business, modeling against those, geospatial to be super hyper efficient with those media dollars,
and then an incrementality test to prove it out.
Where could your next growth opportunity come from?
Talk with us about your growth priorities and how marketing measurement could help you decide where to invest next.
Talk to an M-Squared expert
