Follow Datanami:
February 2, 2015

The Data Foundation Driving YPlan’s Growth

Richard Mapes

The effective access and use of data company-wide is paramount for a company like us. At YPlan we supply a “going out” app that enables users to find and book events on short notice. We strive to provide the best content in every city so users wondering what to do at any given moment will find a variety of compelling events or outings from which to choose.

Because we use data on a daily basis, every employee must be able to ask their own questions and access analytics without help from our internal data team. We required a tool that was easy to use, but could still handle the large stores of data we process on a regular basis.

Many solutions use SQL, but few employees know SQL or have the time to learn it. With a small data team and limited resources, it’s crucial that every YPlan employee has access to as much data as they need, when they need it.

In Search of a Seamless Architecture

With a complex set of data and the need to access it, our architecture is set up with several sources: in-app events data, operational data, and manually maintained data.

  • In-app events data is data we receive from user activities via MixPanel that we store in a Redshift database. The data is then compressed down to summary data and is exposed via a MySQL data warehouse.
  • Our operational data comes in from a backend database that is replicated to another MySQL database and sits next to the compressed, summarized, in-app data.
    YPlan hosts a "going out" app for major cities around the world

    YPlan hosts a “going out” app for major cities around the world

  • Our manually maintained data is used to represent ad-hoc system knowledge, such as filtering out test events or identifying particular events that need special attention.

All of the different data is then collapsed onto a self-service business intelligence platform so that employees don’t have to deal with a complex architecture, but only an intuitive interface representing the different company models and the particular entities with which they are dealing.

Planning for the Right Solution

As our data infrastructure aggregates data from three separate databases, it didn’t make sense for us to again spend time to extract data to a separate database so we could then analyze it. Scrubbing each database or re-treading our infrastructure would prove too laborious and costly. We turned to Looker, which serves as an overlay to our data infrastructure without re-locating or re-loading the data.

Looker is simple to use and has the analytics capabilities to allow our business users to explore data in what we classify as two modes: operational and exploratory. In operational mode, we use data to track known metrics, such as event performance or acquisition ROI, and use it to answer day-to-day questions like which popular events should we tell users about or how can we best reach our target market.

In exploratory mode, we look for insights that might change the direction of the company: do users go out near where they live or work, do different areas have more affinity with particular event types and are there any areas where the potential audience is looking for an event but is poorly served.

Reaping Results

Since the implementation of Looker, we’ve transformed how data is accessed and consumed across the organization. The executive team monitors and tracks how the company is doing through data, like revenue and conversion rates. The looker logo_1marketing team monitors acquisition channels and downloads to see how those are converting into sales. The sales teams gain insight into what events are selling, what we should keep running and how that could affect the general pipeline. The product teams analyze what effect each feature has on user experience and how we can change customer interaction depending on the features.

Analytics and insights are so ubiquitous that, if we were to lose them tomorrow, we would face immediate problems. We would go back to the days where our small data team would have to spend copious amounts of time handling every data request from employees, lengthening cycles and leaving no time to gather additional data or improve operational efficiency. For a company with limited resources, that is an unacceptable scenario when tools exist that enable employees to process data on their own.

One of the most useful insights Looker provided us was when we were exploring the relationship between the number of tickets booked and the total number of bookings via the exploratory mode. Our initial findings showed that most bookings were made for one or two tickets, suggesting that not a lot of people were going out in large groups. While this immediate insight might have led us to focus on small groups, this didn’t match our expectations of people planning an outing on a Friday night, prompting us to look further.

With a little digging, a striking interaction between ticket number and time to book was revealed – while those booking one or two tickets booked fairly soon after first finding an event, the time to book went up dramatically with each new ticket required, jumping from hours to days. A lot of events we sell at YPlan are for the near future, so taking even a few days to book a ticket can result in tickets selling out or even missing the event entirely.

We quickly realized that the problem was in the communication between customers and their friends, with people missing out on events due to waiting on that last person to answer yes or no before buying tickets. What we needed was functionality to allow users to invite friends and keep everyone up to date as to who was going and provide a simple mechanism to prompt users to buy tickets if either time or tickets start running out. This feature is now available in our latest release.

Get More from Your Data

From predictive analytics to data discovery, data is only as useful as the information that can be gleaned from it. Selecting the right data analytics tools for your YPlanbusiness is of utmost importance. We needed a simple, intuitive, self-service tool that allows our employees from all areas to access and interpret data. With Looker, our business insights and the ROI on our data are compounded when employees’ questions can be answered without requiring a lengthy back-and-forth process between departments and the data team. Additionally, liberating data teams to tackle long-term tasks versus everyday general requests will speak for itself.

By smartly employing a data solution, you can ensure that your business operations are streamlined to produce tangible gains. New technology and an increasingly connected world mean capricious trends with hurried consumers. Don’t let that be a handicap to your business; instead tap in to the data being produced to keep pace with your target market.

About the author:  From an initial start as a neuroscientist, Richard went on to develop software for numerous companies such as MetaDirectory for Sun, AdSense for Google and the Data Analysis Pipeline for Qubit. Richard is currently Director of Data Science and Analytics at YPlan, the going out app, where he leads the team to support data capture, reporting, analysis and personalisation.

Related Items:

How NoSQL Drives Analytic Agility at Nielsen

9 Places to Get Big Data Now

Who’s Buying Into Big Data? It’s Not Who You Might Think

Datanami