Big data

What's Ahead for Spanish Wine + Data: Perspectives on Innovation, from Madrid [Bilingual Edition]


This week I’ve asked Andrés Bonet Merten, our lead in Madrid for Enolytics Spain, to offer his perspective on the year in wine + data for that international market. Andrés has been working tirelessly to advocate for the use and benefits of data-driven insights, originating both within a wine business’ own operations and “without,” that is, everyday consumer behavior and sentiment around the business’ wines.

An obvious benefit of having a local market presence for Enolytics is being able to communicate our work in the local language. A second benefit is a more nuanced perspective of the market. Yes, wine + data in Spain is still wine + data, but with a Spanish accent.

Here is Andrés’ perspective on how that looks and sounds, now and for the future, summarized in eight concise takeaways.

  1. The transformation of wineries will need multicultural and diverse teams, and their common language will be programming.

  2. Here in Spain new young oenologists are raising expectations of Spanish wine, and the industry is vibrant and growing at high levels. Some very clear minded wineries are starting big data projects in vineyards, and market intelligence consultants are busy signing new clients.

  3. Technology will soon cover 100% of Spain’s vineyards and robots will do most of the field work in the near future. AI will change people’s lives and the wine business of Spain – even if most of the wine industry is still getting used to thinking about data or analytics.

  4. Spain will have a great viticultural future due to traditional Spain’s mix of research and creativity. Just as the former President of Google Spain, Isabel Aguilera, said that “going against automatization or against the new technologies is absurd, this is the way it will go,” the future of Spain’s wineries will also have to go that way too.

  5. Wine in the future is a “hybrid” wine. The experience of drinking wine will be enhanced by a hybridization of wine and data-driven technology designed to enhance the experience.

  6. The good enough wine has to die before an extraordinary wine can be born. The wine industry has to overcome “what is” and adopt “what might be.” Connectivity will deliver disruption in wine tasting and consumers will value not only quality but also sustained growth, environmental assessment, ethical sourcing, production and commerce in a holistic business market.

  7. The “new Spanish wineries” will need to manage the ambiguity of going local and going global at the time, while using logical thinking and emotions at the same time. They will develop and retain talent and customers, and they will cooperate rather than compete, creating a stronger business culture overall.

  8. The leader of the winery will need to transform the company from inside to outside, first creating new departments fueled by innovative technology that afterwards spreads into all the areas of the winery. The challenges of the future are already here.

Thank you, as always, for reading.

* * * * *

Esta semana he preguntado a Andrés Bonet Merten, nuestro contacto en Madrid en Enolytics Spain, si nos podía dar una visión de futuro sobre vino + datos en ese mercado internacional. Andrés ha estado trabajando sin descanso abogando por el uso y beneficio de los análisis realizados con big data de las operaciones propias de bodegas y de los comportamientos y percepciones de los consumidores de vino.

La ventaja clara de poseer esta presencia en el mercado local español, es sin duda poder comunicar nuestro trabajo perfectamente en su lengua local. Una segunda ventaja es una percepción más fina de este mercado. Sí, vino + datos en España siguen siendo vino + datos, pero con acento español.

A continuación muestro la perspectiva que dibuja Andrés sobre cómo pinta y suena ésto, ahora y en el futuro, resumido en ocho breves propuestas.

  1. La transformación de las bodegas necesitará de equipos multiculturales y multidisciplinares; su lenguaje común será la programación.

  2. Jóvenes enólogos españoles crearán grandes expectativas sobre el vino español y la industria del vino seguirá en plena forma y creciendo a ritmo fuerte. Algunas bodegas preclaras más comenzarán con proyectos de recogida de datos en sus viñedos y consultores de inteligencia de mercado seguirán activos desarrollando nuevos clientes.

  3. La tecnología va a cubrir en breve espacio de tiempo los viñedos españoles y robots harán gran parte del trabajo de campo en un futuro no tan lejano. La inteligencia artificial va a cambiar la vida de la gente como también la de las bodegas – incluso aún cuando la mayor parte de la industria del vino aún se está acostumbrando a pensar sobre la existencia de los datos y su análisis.

  4. España va a tener un futuro vitivinícola excepcional gracias a la tradicional conjunción de investigación y desarrollo con su creatividad. Tal y cómo dijo Isabel Aguilera, la ex Directora General de Google España, “ir en contra de la automatización o de las nuevas tecnologías es absurdo… va a ser así”, y en el futuro las bodegas irán por ese mismo camino también.

  5. El vino en el futuro va a ser un “vino híbrido”. La experiencia de beber vino va a ser potenciada y transformada por una hibridación de vino con una tecnología alimentada por datos diseñada a tal efecto.

  6. El vino “suficientemente bueno” va a tener que morir para que nazca el vino “extraordinario”. La industria del vino ha de superar “lo que es” para adoptar “lo que podría ser”. La conectividad nos ofrecerá la disrupción entorno al consumo de vino y los consumidores de vino no solo valorarán la calidad, sino también el desarrollo sostenible, la evaluación medioambiental, el aprovisionamiento, producción y comercialización bajo parámetros éticos en un mercado del vino holístico con múltiples interacciones.

  7. Las “nuevas bodegas españolas” van a tener que gestionar la ambigüedad de desarrollarse localmente como internacionalmente a la vez, la de utilizar la lógica y las emociones conjuntamente. Van a tener que adquirir y retener talento y clientes a la vez, cooperando más que compitiendo y desarrollando una cultura de empresa más sólida adoptada por toda la organización.

  8. El líder de la bodega va a tener que transformar la compañía desde dentro hacia fuera, creando primero nuevos departamentos impulsados por una tecnología innovadora que posteriormente se extenderá a todas las áreas de la bodega. Pero éstos retos para el futuro ya han llegado, y han llegado para quedarse.

Gracias, como siempre, por vuestra lectura.

How to Work with Sales Data from Distributors: A Case Study from Paso Robles

Jason Haas head shot.jpg

This week we’re shifting gears away from consumer data and looking squarely at data that’s internal to the winery itself.

Inspiration for this post comes from Jason Haas (above), Partner and GM at Tablas Creek Vineyard in Paso Robles, who wrote in December about not being able to evaluate what you don’t measure.

“I hate it when I feel that the data that we’re capturing doesn’t represent the critical decisions that customers make,” Haas wrote on the Tablas Creek blog. “Because that’s the important thing about data: It lets you know, beyond anything anecdotal, whether you’re doing a great job or not.”

To which we say, Word.

Haas uses data in many ways to measure the effectiveness of their sales channels, including:

  • Relative effectiveness of different sorts of visitor tasting experiences, such as standard tasting versus reserve tasting
  • Measurements of value and happiness of wine club members, such as median length of membership and average additional sales to members
  • The effectiveness of offers and promotions
  • Engagement, meaning the percentage of people who open, click on or respond to emails.
  • Haas also delves into publicly available databases for insights, as he did in this blog about ratings of unfashionable grapes and this blog about how people age Tablas Creek wines.

And then there is the sales data that Haas requests from distributors who represent his wines.

To an outsider, this seems like a direct enough request – to be informed about where and how much of your own wine is being sold. But responses and results that Haas receives from the distributors is mixed.

“Some distributors are great and can (and are willing to) give you data in the format and depth you want,” Haas said. “Others are constrained by systems that don’t export data usefully, or can’t be automated to do so, and so you have to ask for it each time you want it, which with 60+ distributors is a nightmare.”

Yet he persists, with the goal of answering the following questions:

  • What wines are selling?
  • Where are they selling them, and how broadly? Were those 30 cases sold to 15 restaurants, say, or to two retailers?
  • What wines are being sampled by their salespeople?
  • What is in inventory?
  • What is the current pricing and what incentives/deals are being offered?

Haas said that, generally speaking, the more important they are to a distributor, and the more nimble and independent the distributor, the better the data and responsiveness to the request.

Most of the time, there is nothing actionable in the data he receives on a weekly or monthly basis. But he does try to look more comprehensively mid-year and at end-of-year, at the states for which he gets good data.

"If I see something that worries me I reach out to our brand manager to make sure he or she is aware of what’s happening, and maybe as important, aware that I’m aware of what’s happening," Haas said. "It’s not a guarantee of things improving in the way you want, but it dramatically increases the chances."

Haas is able to be more detailed when it comes to sales reports within California, taking these three steps:

  1. He sorts the sales data by region to determine where their market work will be most valuable in the coming months.
  2. He sends Thank You notes to reps or managers who are doing a good job.
  3. He re-sorts the data by item to determine any trends. Is a wine selling faster (or slower) than it has been? If so, why? Did we lose one or two high velocity placements? Or is there something broader-based going on? Should he ask the distributor to distribute some samples of a particular wine to a team or a group of teams? Or should he call the brand manager to come up with a more creative incentive program to help improve focus?

"None of these things are magic bullets in sales," Haas said. "It’s a crowded marketplace and there are lots of other smart people out there competing for your business. But knowing what’s actually happening in your key markets and communicating what you’d like to see to your distributors – who, after all, have the same goal as you, to sell your wine – definitely improves your odds."

Is your team already taking these steps? If so, I'd love to hear about it.

If not, how can we help move you in the right direction?

I look forward as always to your thoughts, and thank you for reading.

How to Visualize Wine Consumer Data


Let me show you a picture.

What you see above is a snapshot taken from a project we did recently for a California winery, who wanted to know how their brand was performing within the eyes of consumers in major markets within the U.S.

(We love this kind of question. It's answerable, first and foremost, using quantitative data of hundreds of thousands of data records that consumers spontaneously and objectively generated themselves.)

We took that data, packaged it, and built an online interface for the winery that they, and our team of analysts, "sliced and diced" according to the questions that the winery wanted to know.

The pie chart, above, is a screenshot of an answer to one of their questions, which was how their brand was performing in major markets.

So we showed them their brand's top ten markets according to consumers.

We showed them the share of consumer interest that their brand owns in each of those markets. New York and San Francisco, sure, but Denver and Cleveland? Why not Boston or Atlanta? And why is Washington DC only #9?

We showed them how consumers were talking about their wine within the different markets (which helps to explain the top ten line up).

We showed them how their share of consumer interest in each market evolved over the past four years -- which markets grew, which markets dropped off the list, etc.

We showed them how, if they click on any of those slices of the pie, they could see heatmaps that indicate exactly where consumers were located in each market at the time that they generated the data record about this particular brand.

And then we showed them their competitors, that is, other wines within the same price category that were also competing for consumers' interest. Sometimes they had specific labels in mind and they wanted to see their own performance compared to those labels, and we looked at them too.

They said, "Cool."

We said, "Cool. What else would you like to know?"

And on it went.

Here's the bottom line: This is information that wineries can use in order to understand consumer interest. It's information that's responsive to unique, direct questions from the winery itself. It's information that is helpful for communicating more effectively to consumers, which means selling more wine.

Can we help you with that too?

Please let us know.

Thank you.

How Data Tells the Story of Wine

Summer! Officially!

Time for long weekends, beach reads, and cool, refreshing drinks in our glass.

So it totally “fit” to be working on a project this week about sparkling wine.

The idea was to get a baseline understanding of consumer behavior around sparkling wine. By consumer behavior, we mean things like the categories that consumers are rating highly; the occasions that consumers are celebrating, and their everyday sentiment toward bubbles too; and the performance of Prosecco versus Champagne versus Cava, as well as the best-performing markets for each.

It’s been a lot of fun, and incredibly interesting too to dig into the details of those queries.

What also became clear is how much more can be done.

One thing we’re beginning to see is the value that non-wine-focused data can also add to a project like this.

It’s like doing the research to write a compelling story. As we develop the narrative to get us to the end goal, we see what other sources of information can help to flesh out the plotline. Socio-economic data, for example, tells us things like what kind of house the consumer lives in and what sort of income they’re bringing home. Detailed geo-located weather data describes the scenery and atmosphere around a consumer’s behavior at a specific time of year, like a hot, sunny holiday weekend when searches and scans spike for sparkling wine. And etc.

The point is that these data streams help to sketch in the profile of the consumer, and they provide the context for their behavior involving particular styles of wine.

Data might seem very binary and very zeros-and-ones. But there’s a tremendous amount of creativity involved, too, in using the data to create the narrative that answers the questions the client needs to know.

Enjoy this long Memorial Day weekend, and thank you as always for reading –

Direct To Consumer Dashboard Drivethrough, Part Two: SKUs


Last week we took a first look at an interactive dashboard for the big data that a wine business already has, specifically from its Direct To Consumer (DTC) program.

The DTC program is top-of-mind for just about every winery I know. We wanted to demonstrate how interactive dashboards can help to identify problem areas and work toward optimizing DTC's profitability.

We looked at the value that is released when data is visualized, accessed, and "sliced and diced" in real time. We started with KPIs -- things like cases sold and month-to-date and year-to-date comparisons -- and how clicking on any one of the KPIs presents a fresh, detailed visualization of variables that matter most to the winery.

This week we're turning to a different set of visualizations, namely underperforming SKUs based on low margins and/or low sales. When you have hundreds of SKUs, how do you compare their performance relative to each other and in an absolute way, while also still allowing to slice and dice by channel, wine type, etc.?

One visualization option is to use a scatter plot (above). It displays all the SKUs, using your selected filters, on a chart by margin and by revenue. This makes it easy to identify your top performers -- those with high margins and high revenue -- and the bottom performers, with low margins and low revenue. Those are the SKUs that need attention and, possibly, a plan of action. With Enolytics' interactive dashboard, it is simple to filter out the bottom performers that are not meeting your revenue and/or margin targets by using a simple lasso tool.

The lasso zooms in on a target area, which results in a focused data set that is useful to further analyze. The list of bottom performers is also displayed in table form for easy understanding and selection.

Enolytics' dashboard enables you to immediately investigate, based on objective insights, in order to facilitate business decisions to correct performance.

Bottom line: improving your understanding multi-fold, particularly of your biggest challenges, weaknesses, and opportunities. 

Can we help you with this? I'd be very glad to talk it through. Give me a call (+1.702.528.3717) or send me an email ( anytime.

Thank you as always for reading --

Wines that Consumers Want, versus Wines that Restaurants List

This week I'd like to bring your attention to a surprising pattern.

We noticed it as we've been working to integrate fresh big data sources about wine consumers in the on-premise and mobile environments.

It's more than a pattern, actually. It's a discrepancy and a significant one, between these two things:

What consumers want, compared to what restaurants are offering.

There's a significantly bigger gap than what we expected to see, whether we're talking about specific varietals or particular styles.

We have access to tremendous amounts of data on wine consumer behaviors, yet insights from analysis of that data aren't being "translated" onto wine lists to nearly the extent possible.

Why is that?

A number of possibilities come to mind, but I'll never know for sure which one is accurate.

Here's what we do know: the discrepancy between what consumers want and what restaurants are offering is risky business.

Unnecessarily so, given the depth of sources and the breadth of insights available.

The intelligence is out there. It's a question of accessing it.

We can help.

What discrepancies or questions do you need to check, and address?

I'd be glad to hear about it. 

PS This past week I had the pleasure to be interviewed by Lee Schneider of Red Cup Agency, on a podcast about Enolytics for their fantastic cult tech podcast series. Producing a high-quality podcast is a lot tougher than it sounds, and I'm grateful to Lee for the chance to talk about our work. I hope you'll take a few minutes to tune in.

Thank you for reading, and for listening -- 

How Do I Use My Data? Let a Data Hero Demonstrate.

Here’s something we hear a lot at Enolytics:

We think we’ve got some good data, and we’re pretty sure it’s valuable. We just don’t know what to DO with it.”

It’s a fair thing to say. That’s partly because there aren’t a ton of people in the wine industry who work with data, so talking about it (and familiarity with it) is far from common. It’s a fair thing to say, also, because each situation has a lot of moving parts – your data will be different than another company’s data and you will each need it to do different things in order to be useful for achieving your company’s individual goals.

So I’d like to spend a few moments in this space to share some stories of Data Heroes. They’re the people who defined a goal, accessed the right data sets, and asked the right questions of that data in order to derive the most useful insights that they could put into practice to achieve that goal.

Today’s Data Hero comes from the entertainment industry, where the production company behind the Godzilla movie used data to evaluate everything from the trailer to cast selection to the release date. The company is called Legendary Pictures and their analytics office is based in Boston, far away from Hollywood groupthink.

Here’s an important takeaway from the Godzilla case study: Legendary used the data before they filmed and launched the movie, so that they could be as confident as possible in its success. Much of the work we see with data in the wine industry comes after the fact, but that’s the difference between reactive or diagnostic analytics (when we look at what happened) and prescriptive analytics (when we look at what we can make happen).

Here are some additional takeaways from a Data Hero that took the prescriptive analytics approach:

  • Legendary combined both new and traditional data sources, such as social media engagement and hourly box office figures.
  • Legendary learned to spend fewer marketing dollars on hard core fans and more on “persuadable people.”
  • They used data to identify those “persuadable people” and, more importantly, they segmented those people into precisely defined cohorts and communicated messages targeted precisely to them. Cohorts consisted of as few as four people each.
  • Legendary engaged and listened to their audience, and crafted the movie’s trailer to align with the audience’s interests. In practical terms, that meant the trailer focused more on the leading actor and conspiracy theories (which the audience loved) and less on the monster and destruction (which the audience did not love).

The transfer of each of these takeaways to applications in the wine world is not only possible. It’s the way forward, and it's the shape of the model that Enolytics is building. Combining new and traditional data sources. Allotting marketing spend to “persuadable people." Defining precisely who those persuadable people are. And crafting communications specifically to them.

Do you know of some Data Hero cases in the wine industry? Please share them in the comments section or with us directly via email at

Click here for more information on the Legendary Pictures case study.