May 2018 Issue of Wines & Vines

How Wineries Take Advantage of Big Data (or Any Data)

by Andy Starr

Are the winemaking and sales decisions in your winery data-driven? Are all of your data sources well integrated into one system? Ask the question of some winery owners, and you may get these responses:

• "I will use data to run my winery once a really well-known Napa Valley winemaker does it and gets a98-point score. It's the wine industry, where everyone strives to be the first one to be the second one."

•"My winery prides itself on making wine exactly the way they did in France in the 18th century. I evenhave my tasting room staff wear Les Miserables costumes at work, and on weekends when we have bigcrowds, I make them sell their teeth just like Fantine did. The tourists love it!"

• "Data analysis is what the 'big guys' do, but not here. We're just farmers."

• "Wine is an art."

True, I invented the quotations above, but they're not as far-fetched as you might think. Data is generated at every step in the winemaking process, from soil moisture to vineyard sampling to weight tags to all steps of winemaking to case goods to sales to consumer data. We are buried in data, yet we often make important decisions on anecdotes and "gut instinct." Everyone has a story of an owner saying more or less, "I love our Gold Digger's Reserve, named after my last three wives. Sure, we only sell 173 cases every year, but let's keep making 1,000 cases."

Marshall Graves is a vice president and wine industry specialist at Bank of Marin. He encourages wineries to "go beyond using your data set to simply resell to existing customers. Aggregate and examine your data to obtain a precise understanding of what your club members enjoy beyond wine."

Graves suggests accomplishing this by collaborating with companies that specialize in data analytics and social physics. "Done well, this will improve wine club retention and ultimately result in wineries having a more efficient and cost-effective marketing strategy." He added that "each of the top five wine-producing states also have established technology hubs. Those who engage in conversation with their tech community will become agents of change vs. being the disrupted."

We all know we need to use data better, but turning it into something that objectively drives decision-making is a challenge for nearly everyone. I interviewed representatives of two wineries and their data vendors who are successfully integrating and using data in their everyday operations.

DtC and consumer data drive decisions

Michael Longerbeam is the direct-to-consumer manager at Dry Creek Vineyard, a 100,000-case operation in Northern Sonoma County, Calif.. Founded in 1972 and still family-owned, Dry Creek relies heavily on direct sales. Longerbeam is a marketing veteran, having done it in both software and wine, including Windsor Vineyards, one of the industry's DtC pioneers.

Longerbeam observes that most people see the wine industry as a relational business, where decisions are often made on interpersonal contacts. He encourages including data analysis in addition to conversations with industry colleagues when making business decisions. Longerbeam suggests looking at three things in your DtC program:

  • Recency. When was a customer's most recent purchase?

  • Frequency. How often does the customer purchase?

  • Monetary. How much does he or she spend per purchase or per year?

Longerbeam used to do all the data crunching himself, joking that the other DtC staff see him as "the nerd on the computer." He creates and slogs through a number of spreadsheets, with the intent of distilling the data down to something useful or actionable. He warns that doing it on your own, you can "go snow-blind by all the data," and not end up with the result you want.

His most relatable example is using data analysis to increase wine club retention. His team had discussed giving every club member a gift to improve retention. That may help sales, but it's expensive. Sending no one a gift costs nothing, but you may lose some revenue and profit from club drops. By grinding through the data, they learned there was a spike in cancellations at a certain number of months. They then used that data to send gifts to only those club members who had been in the club that long, showing them some love at a point when they were most likely to break up. Longerbeam then measured club retention, and it improved. He notes that analysis "can get granular, but that's where the power of the data is."

Another example was deciding on what their next wine should be. My observation is most wineries make this decision based on something like "the winery down the road makes that varietal and it sold well" (another example of the "first to be second" rule).

Instead, Dry Creek worked with Enolytics, a wine data analysis company. Enolytics built an interactive dashboard using Dry Creek's existing, anonymized data, and determined that the winery was missing a wine at a certain price point. That data could be used in conjunction with other factors like wine quality, terroir, etc. when deciding on new wines. Longerbeam's observation is that "Enolytics is visually driven, making it much easier to find the data you seek."

Something everyone should have
Enolytics produces for Dry Creek something that everyone should have-a breakout of their wine portfolio by margin and volume, displayed in an easy-to-understand graphic. "This forces you to look at your portfolio," Longerbeam said.

Enolytics founder Cathy Huyghe (pronounced HOY-huh) created a company devoted to giving a winery a clear picture of the wine consumer. Typically starting with a winery's existing DtC data, Enolytics can add data from a range of sources including Vivino, which gathers data from its 29 million-plus downloads. The data can be segmented by brand, varietal, price point, appellation, etc. In addition, Vivino data records include a latitude and a longitude, which provides visualization into consumer location and demographics.

Huyghe gave some examples of how the company's data-gathering and analysis arms a sales team with "objective, quantitative evidence generated by consumers" instead of typical reliance on a distributor rep's qualitative and anecdotal assessment. Huyghe noted this causes data analysis to move from "hindsight to foresight." Two examples stood out:

Using its range of data sources and geo-tagging, Enolytics could create a "heat map" showing which neighborhoods within Manhattan were frequently chatting about Russian River Pinot Noir. Huyghe offered that the map "informs you where the fish are, so you know where to go fishing." A winery could then instruct its sales staff and local distributor to focus their sales efforts in those neighborhoods.

For other clients, Enolytics has been able to identify subregions, e.g. suburbs, where a brand's consumer interest is currently strong. The company then finds the additional suburbs around the country with similar demographics. Those become the next target markets.

Data integration from grapes to consumer

Zachary Rasmuson is the senior vice president and chief operating officer of Duckhorn Wine Co., a collection of premium brands and wineries that includes Duckhorn Vineyards, Paraduxx, Decoy, Goldeneye, Migration and Canvasback. Rasmuson started as a cellar rat at a Napa Valley winery, then went to Goldeneye in Anderson Valley, and worked his way up to leading all operations for the parent company. He takes pride in building teams, systems and infrastructure.

Duckhorn Wine Co. is experiencing rapid growth among its six wineries, creating management challenges for planning. While Rasmuson notes that sometimes "harnessing winery data is just inventory management," collecting and organizing all the data each year is seldom as straightforward as he'd like.

Combined, the six wineries produce 130 distinct wine SKUs. How does the staff know they will have the correct amount of fruit assigned to each of them? How many new barrels do they need? What is the cost of each wine? To answer these questions, Duckhorn turned to Oztera, whose Teravina integrated data platform combines the "data silos" of production, lab analyses, finance, inventory, wholesale sales and DtC sales into one system.

Now Rasmuson can easily see and address potential fruit shortages for a particular wine and know how many barrels to buy. "The brilliance of Oztera is that it communicates well to finance and accounting, from cost of fruit and barrels to capital expenditures. It's a nice, nifty platform," he said.

Trial blend platform
Oztera was able to create a standardized trial blend platform for Duckhorn, which replaced the silo of each winemaker creating and using his/her own spreadsheet that resided on their personal computers instead of the company database.

Ultimately, Rasmuson said that data collection is often most valuable in the vineyard with their weather stations, soil monitors and other probes, and with their consumers. Data for winemaking decisions can only go so far, as the winemaker's skill is "to adapt to a current vintage in real time, especially in the first weeks after harvest." Rasmuson is intrigued by data that could be used for repeatable winemaking results, such as phenolics assays where winemakers fine-tune their palates to a phenolics reading, then try to match the phenolics from vintage to vintage.
 I interviewed Mike Stallman, Oztera's Chief Geek Juggler (yes, that's his job title). He explained that Oztera, based in Silicon Valley with offices in Napa, Hawaii and Arizona, specializes in solutions built on the Microsoft Dynamics NAV platform to address the unique challenges faced by wine companies. Oztera's Teravina program leverages Microsoft's investment and development of a robust system used by 140,000 other companies in a wide range of businesses.
 Stallman notes that a winery typically uses many separate data systems for winemaking, accounting, sales, etc., creating numerous business challenges. Data is transferred back and forth between systems periodically, a clumsy exercise that invites a "spider web of problems." The Teravina system integrates all these systems into one.

As an example, many wineries struggle with product costing. With a typical siloed accounting program, wine costing becomes an exercise where the finance manager builds multiple Excel spreadsheets, bugs winemakers for their winemaking data, and then locks her/himself in a closet for two days to figure it out. Instead, Teravina integrates grape contracts into its winemaking and accounting programs. The system can compare contracts with weight tags and then follow the wine as it moves around the winery, adding in packaging costs and overheads. By integrating the entire process, everyone knows the true cost of their wine.

Stallman believes the Teravina system helps maintain quality consistency from year to year, by giving winemakers the ability to compare blending and tasting records, plus other historical production data. "We provide winemakers the tools to let them do their job more easily."

Ultimately, Stallman sees that an integrated system provides "one source of truth. With multiple systems, you can get more than one truth." (Or multiple sources of half-truths.) Once you have all the data in one system, "you can ask just about any question, and we can build something to answer it. You can see everything from grapes to bottling projections, plus report on your KPIs within one site." Many clients use Oztera's system to plan their bottling runs over the year, and bring packaging vendors in on the process.
Oztera claims its software platform is simple and cost-effective for small producers, who can realize the same benefits as the big guys. Their customers range from an Iowa boutique winery producing 15,000 cases to a big California operation making 3 million cases.

Can you afford it?
You may be reading this and thinking you can't afford it. I would reply with, can you afford not to? You should be able to find someone who can help, as tech research hubs are located near winegrowing regions. If the cost of hiring a data analysis service provider is too high, ask your local wine association to subscribe and get a group rate for its membership. Perhaps Dry Creek's Longerbeam summed it best. He would like to see the wine industry use data the way it is already used in high-tech and other consumer goods companies. "You're sending emails to your customers anyway, you might as well learn from it. Next time you make a decision, work data into it."

Andy Starr, founder of StarrGreen (, is an entrepreneur, marketing manager and winemaker who provides strategy, management and business development consulting services. A resident of Napa Valley, Calif., he holds a bachelor's degree in fermentation science from the University of California, Davis, and an MBA from UCLA.

Print this page   PRINTER-FRIENDLY VERSION   »
E-mail this article   E-MAIL THIS ARTICLE   »
Posted on 05.12.2018 - 08:47:03 PST
Tremendously valuable article. Thank you. The coming together of disparate pieces of data into a single analytics platform for better and faster decision making is the missing piece the wine industry can use to optimize operations and expand. Enolytics adds great value in turning data into powerful usable information. Consumer sensory and attitudinal data from Quini (QUINI DATA™) delivered in near real-time in the form of ready to use analytics and actionable insights customized for each executive will cover the industry's 'last mile' of missing knowledge. This data, coming from the ultimate decision maker, the consumer, answers the 'why' that today eludes marketing, sales, R&D and C-level wine executives. The wine industry is at the early stage of exploring data, but the opportunity to leverage data is immense.
Roger Noujeim

Posted on 05.16.2018 - 12:22:06 PST
Andy - good article and starting point for data analytics, and you're just scratching the surface. You really don't tap into "Big Data" analyses here, not get to the full level of club data digging. As for affordability, the ROI of these efforts is typically 5X or greater, if paired with effective marketing campaigns to reap the benefits. Miles to go here...
Joel Miller