Vivino leverages data to drive growth and the best possible customer experience with Looker.

Case Study: Looker & Vivino

Vivino is the world’s largest online wine marketplace and most downloaded wine app, powered by a community of millions. Vivino’s unique wine shopping experience uses community data to provide personalized wine recommendations—making wine discovery and purchase fun, accessible, and effortless for wine drinkers of every level. Vivino is experiencing fast year-over-year growth by using data to fine-tune retention strategies, improve customer service and response, and deliver data-rich partner apps for winery and merchant partners.

With company-wide access to data, Vivino is able to deliver the best information and deals on wine, while also serving up the best possible customer experience. As they expand their use of data to deliver insights to their winery and merchant partners, they’ll bring the industry never before seen worldwide visibility to wine.



Key Takeaways

“It was a big concern of mine because it’s important that people trust data, and, if you have two high-level people on different teams presenting different data to the executive team, then they start to always question the data.”

A proliferation of spreadsheets spurs a search for an alternative

William Moor, now director of business intelligence, joined Vivino in 2015.
At the time, the majority of the small company was based in Copenhagen while William was one of the few employees in San Francisco. To serve the entire organization’s data needs, William spent most of his time running a massive amount of queries for non-technical users and trying to make sense out of a plethora of spreadsheets owned by different stakeholders—most of which were in different timezones.

In addition to helping non-technical users, Vivino’s technical users who ran their own database queries also needed help making sense of conflicting results from queries run without consistent data definitions across the organization.

Since everyone had their own data sets and metrics, conclusions didn’t match up as they should. The data was inconsistent. This was particularly problematic on the eCommerce side of the business, where data—whether correct or not—was being used to influence strategy and decision-making as the Vivino team scaled globally.

“It was a big concern of mine because it’s important that people trust data, and, if you have two high-level people on different teams presenting different data to the executive team, then they start to always question the data,” Moor adds.


of the Vivino team uses data daily to make informed decisions


increase in customer retention


improvement in Net Promoter Score (NPS)

A unified way to access a single source of truth

When Moor embarked on his search for a better solution in 2018, his goal was to find a flexible and scalable solution that would enable organization-wide access to trusted metrics as their global team (and data volume) began to grow.

After evaluating different solutions, Moor chose Looker for its ability to provide trusted metrics across the organization, which he knew would help every department make better decisions for their customers.

With LookML, Looker’s modeling layer, he knew he could ensure data governance and metric definitions across datasets and queries, bringing trust back into the data and decision-making process.

“I love the ease of LookML. If you understand coding, you can pick it up instantly. I knew it was something that I could take care of on my own and then easily teach the team as we scaled,” Moor remarks.

Utilizing Looker’s native connection with Amazon Web Services (AWS) Redshift, Vivino’s data warehouse, the data team was able to leverage Redshift’s processing power while ensuring a single source of truth for all the queries that ran against it.

By storing all of their data in AWS Redshift and analyzing it in Looker, the Vivino team is now able to view data generated by Mixpanel and Heap (product analytics and event tracking tools) and Salesforce (customer relationship management
application) all in one place, providing a clear and trusted view of the entire business.

Optimizing orders to ensure happy customers

Having instant access to trusted data has been especially helpful to the eCommerce and operations teams at Vivino—enabling Vivino to dramatically scale its marketplace business that is experiencing fast-paced year-over-year growth, with over 12 million wines, 44 million users, and 1 billion scans as of early 2020.

In addition to using Looker to track wine orders and shipments to customers to understand high-level trends in delivery time and satisfaction, the eCommerce team also has custom alerts set so they can be notified if an order fulfillment issue arises, relating to the accuracy or timely delivery from their winery partners. As a result, the entire Vivino team can identify and respond to issues quickly, and then discover and standardize processes that work best. The team uses dashboards, explores, and alerts in Looker to quickly identify and take action on orders that have been in the queue for too long. This ensures a positive and timely customer experience, helps the team stay efficient, and provides peace of mind so if orders are being delayed the team will know about it and be able to take action and find a solution.

The visibility Looker provides has been especially helpful in identifying unhappy customers and providing insights the Vivino team can use to improve and develop retention strategies, and ultimately improve the overall customer experience.

Before Looker, like many eCommerce businesses, the Vivino team didn’t have much visibility into customers who may have had a less than positive experience. Generally, customers won’t tell online merchants why they are dissatisfied—instead, they simply walk away and never interact with them again.

In the past, a customer may have received a shipment of wine they did not like and then never placed another order again. With access to Looker, the team can now identify dissatisfied customers, reach out to them quickly to address their concerns, and ultimately increase customer retention. Using one dashboard, the eCommerce team is able to connect how users rate wines after they’ve placed an order.

This provides a more complete picture, understanding not only the ordering process but also the experience of receiving and enjoying (or not) a particular wine. This team has custom alerts set up from this dashboard to notify them when a customer provides a negative order for a bottle of wine in their shipment. The operations and customer service teams are then notified, and they immediately take action by following up with the customer.

“Our marketplace offering is evolving and, as our needs change, our team can be dynamic and adjust—with the help of Looker.”

Easy access to sentiment analytics
enhances customer support

The customer service team is continually gathering as much customer feedback as possible to help enhance the quality of service and products they offer customers.

One of the metrics they use to track this is Net Promoter Score (NPS), a customer loyalty metric that assesses to what extent a customer would recommend Vivino, its product, or services to friends and family.

Prior to using Looker, Vivino’s customer survey results were distributed across multiple third-party platforms and regional tools, making it challenging to collate, manage, and gain insights from customer feedback. By accessing this data, along with their NPS scores and order analytics in Looker they gained a better understanding of the post-purchase and post-delivery customer experience. These insights are available across teams, providing everyone with a consistent view across all platforms, regions, and partners.

With data centralized in Redshift, Vivino is going a step further and using natural language processing (NLP), which takes unstructured data in the form of human language (text) and renders it into a format a computer can understand to enable sentiment analysis on the surveys.

Because Vivino operates in a number of different markets with different languages, NLP analysis can be especially complicated. To make it easier, they use AWS Comprehend, a machine learning-powered NLP tool that enables them to translate text and derive insights. Once the sentiment analysis is completed, the team applies internal customer service classifications (such as “late shipping” or “incorrect order”) to organize the data into “buckets” in Redshift. When the customer service team surfaces this data in Looker, they are assured it’s clean and accurate. They can run a report and look for key terms, different data entities, and sentiment analytics to understand what customers are saying.

The result? An increase in Vivino’s NPS score (which was already high) by 30%. This also enabled Vivino to expand its team to spend more time tracking and acting on insights from Looker and supporting customers rather than sifting through multiple spreadsheets with inaccurate data.

Other departments at Vivino are also taking advantage of the combined power of Looker and Redshift. The customer relationship management team, for example, uses marketing data from multiple sources that reside in Redshift and is visualized in Looker to identify the marketing activities that increase customer happiness. From there, they can determine what works best across regions to increase customer acquisition.

The analytics team members are avid users of the easy-to-use Looker application programming interface (API). Rather than managing notebooks with queries, they make use of the API to build the data sets they want, run queries, schedule reports, and more.

From data silos to data culture

Since implementing Looker across the organization, Vivino is able to provide company-wide access to trusted, accurate insights that help everyone make the best decisions for their customers. With the time saved from running manual queries and double-checking accuracy, Moor and his team have been able to go beyond dashboards and integrate insights into employees’ workflows, including setting up custom email alerts and using the Slackbot to bring data into conversations.

“Over half of the company uses Looker at least once per day, and this includes engineering teams. If I walk around the San Francisco office, everyone has Looker open and is doing something with it,” Moor points out.

“Over half of the company uses Looker at least once per day, and this includes engineering teams. If I walk around the San Francisco office, everyone has Looker open and is doing something with it.”

Up next: Embedded analytics for
winery and merchant partners

In addition to using Looker for internal analytics, Moor’s team is also leveraging Powered by Looker to embed analytics into external products for their winery and merchant partners, allowing these partners to gain insights from business intelligence without needing to build out their own data stack or processes.

For years, the Vivino sales team has been using Looker to track metrics for their partners, including how their pricing compares to other merchants, or how customers rate their wine compared to other wineries. It became clear that wineries and merchants were craving data to better understand and build their business plans as they kept coming back to the Vivino team with more and more questions.

They wanted to interact with the data themselves.

“We found that the wineries were super-hungry for data and wanted to be able to answer their own questions: How are users rating my wine? How well is it selling? What are people asking and saying about it? How are my brand pages performing?” relates Moor.

Clearly, this presented a great opportunity for Vivino to build an app,
especially for its partners.

The Vivino product team started thinking about what type of product they should develop and considered whether they should build or buy the analytics portion. When the product team heard Moor’s presentation about embedding Looker into the partner application, they realized immediately that this made sense. Rather than allocate expensive engineering resources to build the application from scratch, they decided that going with Looker again would be the best solution because they were deeply familiar with the tool and could leverage the data model they had already built for their internal analytics.

In addition, since Looker is entirely web-based, it’s simple to rapidly build custom
analysis capabilities.

The product team went ahead with the plan and recently launched new functionality in the app targeted at winery partners, which enables them to benefit from Vivino’s massive database, including data from user reviews on wineries and wines. Previously, Vivino’s profile for each winery was written by the internal team.

Before rolling out the app, Vivino’s product team asked the wineries if they were interested in a product that would enable them to add custom information to their profiles, including videos. The wineries were extremely receptive to that idea, and several are already on board in the early product release.

“Essentially, this is a portal page for wineries to get timely data on how users are interacting with their wines. From a marketing perspective, they will be able to understand their wine ratings and how they stack up against the competition,” Moor explains. “For instance, if a winery makes a Napa Valley Cabernet and sells it for a certain price, they can determine if the price is in line with what the market will bear for that type of product.”

The Vivino team is also evaluating a data product for their wine merchant partners, who use the Vivino app to sell their products. This will positively impact their businesses because they will be able to use embedded Looker dashboards within Vivino to analyze and spot market fluctuations and trends.

Merchants will be able to check their pricing against competitors, compare their stock to their competitors to gauge the uniqueness of their inventory offering, and spot trends in their region early to help them adjust inventory accordingly.

Enough small talk. Let's build something together.