For more than 20 years, OpenTable has connected foodies and novice diners with the restaurants they love. But how does its technology work on the back end? To make a long story short: data.
Beyond the app and website, OpenTable provides restaurants with software that manages their floor plans, phone reservations, walk-ins, shift scheduling, turn times, and more. They can provide all of these crucial metrics to help restaurants run more efficiently, while also providing benchmarking data within their cities and among their competitors.
Recently, Cindi Howson sat down for a Q&A with OpenTable’s Grant Parsamyan, Senior Vice President of Data and Analytics. Together, they talked through OpenTable’s migration from on-premises data centers to the cloud, how OpenTable provides the right metrics for restaurants owners, and how data helps drive better hospitality and greater profitability. Read on for more.
Cindi Howson: Tell us how you've been able to build success metrics to prove the value of OpenTable over time.
Grant Parsamyan: We've tried to measure the incrementality that OpenTable brings to the restaurants. This means people found restaurants through our recommendation engine that they wouldn't have found through another network. We're able to show restaurants the business they otherwise wouldn't have gotten. There is also the efficiency gained from not having a person answering phones, and the convenience for the consumer to be able to make a reservation in just two clicks. We're not focusing on ROI calculations as much, but we have different dashboards and reports that we've created over the years that show the incrementality that's coming from OpenTable to those restaurants.
Let's say you got a thousand reservations from OpenTable: how many of those users were undetermined when they started their search? They were browsing; they were looking at our recommendations and we matched your restaurant to them. We could show, for argument's sake, 30% of those reservations came from undecided diners. That's incremental business.
Cindi: In terms of technology trends, what are you most excited about?
Grant: That it’s less about developing views and more about giving the necessary tools to the end-users to create their views. I think, in the BI space, predefined views are going to become obsolete in a few years. Tools become more sophisticated when we provide users the ingredients and platform where they can build what they want. The more you can shorten the time between someone who has a question about the business and finding the answer, the better it is for the business.
In terms of scalability, I think database technologies that are enabling users to have more elastic processing have been game-changers. In the past, when you had to grow your data storage and processing at the same rate, it was a stopper for folks who had deep data and had to continue buying more processing. I think the new vendors that are decoupling the two are enabling users and companies to increase performance without increasing the costs.
Cindi: Some people argue that you can't do self-service analytics until you upskill users first. Others argue it's like teaching somebody to read. What are your views?
Grant: It's in between the two. When someone joins our sales or account management teams, they go through a data training program where they are taught how to read data. Data literacy is like reading: if you don't know how to read, it’s useless if someone gives you a book. You have to focus on what you want users to learn and then build on that.
Ultimately, everyone wants to do well at their job. People can feel like the investment they're going to put into the tool may not be there. It's the trainer's job to show how they can use data. Data literacy is not a project. It's a program. You constantly have to be doing this.
Cindi: Do you think that historically, the data science industry has spent too much time on technology and not on data?
Grant: For data science teams, it's always about data. Technologies are there to make things easier for us, it shouldn't be the other way around. If you need to train someone on the technology more than on the data, that means perhaps you're not doing your job well enough. Perhaps you haven't selected the proper tool or made the design of things overly complicated. Tools have gotten more segmented in terms of their utility.
I think you should have a portfolio of tools. You don't want to cram every single use case into one single tool because you selected the technology before taking a look at your use cases. I look at the use cases, see what the best solution would be, and then see what technologies would make it easier for the users. It's looking at it from the user's perspective, not from the technologist’s perspective.
Cindi: Can you share how you have positioned the different tools within your portfolio?
Grant: We have several BI tools: ThoughtSpot, MicroStrategy, and Tableau all run on top of Snowflake. The tools can be very overwhelming for someone who is looking for insights. Since ThoughtSpot is great for ad hoc analysis and creating personalized views, we've taken the most used metrics, most common KPIs, and most relevant historical data and put that into ThoughtSpot.
While we would like to put everything in there, there are limitations in terms of the cost benefits. We want to make ThoughtSpot a simple place where any user can run questions without being overwhelmed. Someone can start their journey in ThoughtSpot to narrow in on certain insights.
Cindi: Are you leveraging the data-sharing capabilities on the cloud with Snowflake?
Grant: We started doing data sharing first within our sister companies. Then we began doing a proof of concept by sharing this data with some of our larger restaurant groups. In some cases, we provide turnkey data warehouses for restaurants to use for their own analytics. I can create a virtual data warehouse with their data, and it would serve as an extension of what we already created internally.
As soon as my data gets refreshed, the restaurant group can run turnkey analysis on top of it. More customers are interested in this technology, and data sharing enables that. In the past, you had to monitor extracts, transmissions, and file loadings. Now, it’s simply table-to-table replication. I'm not constrained by volume, so I can transfer terabytes within minutes.
Cindi: Do you see a difference between OpenTable, a digital native company, and your restaurant customers who may not be as data savvy?
Grant: I do, but even digital companies are not fully digital; not all processes are digital. If you still have processes that are offline but are critical components of your business, you're not fully digital. A small restaurant doesn’t have any kind of data besides the reservation book. But if I can tell them more about the diners, I can tell him more about the demand.
Restaurants have traditionally been offline, but more restaurants are becoming digitally savvy. Big restaurant groups specifically are doing a great job seeing how they can increase hospitality through data. Every restaurant wants to provide better hospitality, and how you do so is by knowing more about your diners.
The easiest way to show restaurants the value of data is by showing how they can make more money at the end of the day. You want to make the best decisions to increase hospitality and be the best version of your restaurant.