Mobility is hardly new in enterprise circles. Businesses are busily building mobile apps for in-house use and for customers, while the task of managing employees' mobile devices absorbs many companies' CIOs.
Some organizations, however, look to refine their mobile strategies. When eHarmony reworks its mobile applications, for example, the relationship services provider does so with the backing of a 25TB data warehouse and a roster of data analytics tools.
The dating service has spent the last five years building a business intelligence platform. The data warehouse component employs IBM's Netezza TwinFin appliance and the company's PureData System. A MicroStrategy-based analytics layer sits atop the data warehouse. IBM's SPSS predictive analytics package, the R machine-learning software environment and Informatica data integration software also augment the data warehouse.
The business intelligence tools feed dashboards, keeping eHarmony apprised of site usage and revenue trends. The platform also drives product development at the Santa Monica, Calif.-based company. Data analytics identifies the features and functions users find most attractive on the traditional desktop; those items then can be introduced on eHarmony's mobile apps.
The company's matchmaking service runs on Apple's iPhone and iPad, Android devices and Windows Phone. Half of eHarmony users arrive at the matchmaking service on mobile devices, reflecting eHarmony's extended, social-media fueled reach among younger users. "We're ... working to get the best features and functions from desktop to mobile," COO Armen Avedissian says.
[ More: NoSQL a Perfect Match for eHarmony ]
Better apps means an improved user experience. That translates into higher conversion rates for eHarmony and better matches for customers – and rates have increased more than 25 percent since January. "A successful match is our mission and business," Avedissian says.
Matchmaking Behind The Scenes
You could argue that dating has always been a numbers game, but eHarmony's task is to improve the odds for its customers. The data warehouse and analytics plays a number of behind-the-scenes roles in that regard. One such role is gleaning insight into customers' site usage and level of engagement. The company can keep tabs on users from the time of acquisition all the way to purchase and can also monitor customer lifetime value. It can also perform clickstream analysis to zero in on customer usage patterns.
The warehouse is populated with data from a number of sources, Avedissian notes, including Google Analytics, Adobe SiteCatalyst the SPSS system and the machine learning software. Information is also absorbed from eHarmony's matchmaking Hadoop clusters and the company's transaction database.
The warehouse aggregates that data and runs algorithms that monitor the user effectiveness and behavior on the site, he says. "When we marry all this data together, it really creates visibility."
Product managers at eHarmony monitor their channels and can determine how well different functions and features are doing based on usage. With that knowledge, eHarmony aims to improve its platforms. Successful features thus move from the desktop to mobile platforms.
A new app for Windows Phone, launched in March, represents the most recent example. (The iOS and Android apps were refreshed prior to that.) The mobile apps have inherited desktop traits in a couple of ways. Desktop users, for example, have had the ability to continuously switch their match settings, adjusting such parameters as distances, age and nationality.
Previously, those match settings weren't as finely tuned for eHarmony's mobile apps, but they have since received that upgrade. They can now alter match settings and preferences on mobile and have full dashboard functionality, Avedissian says. Other desktop features, such as an activity feed that lets users see who has viewed and updated their eHarmony profiles, are now available on mobile as well.
Aurimas Adomavicius, president of Devbridge Group, a Chicago-based software design and development company focusing on enterprise mobile and Web solutions, says eHarmony's use of data analytics is common among larger companies with "lots of information or a lot of interaction." He says the volume of data must be sufficient for a company’s data analytics platform to find meaningful patterns.
John Haddad, senior director of big data product marketing at Informatica, sees eHarmony as a good example of a company that is monetizing its data assets. "Web-based companies like eHarmony, Yelp, LinkedIn and Airbnb ... gather massive amounts of data about their customers. If they can't create good data products and user experiences, they're not going to succeed."
Business Intelligence Can Give eHarmony Users a Gentle Nudge
Just as eHarmony can track user behavior, it can use business intelligence to try to influence it. For instance, if a user has only posted two out of a possible eight photos permitted, eHarmony can prompt the user within the app or via email to remind them of the other actions they may take, Avedissian says.
The data warehouse enables that guidance, pushing aggregated metrics and statistics back on to the sites. "Without the data warehouse, we can't put that kind of intelligence back into the websites," he says.
The company also uses business intelligence to monitor the marketing programs that drive users to its products, mobile or otherwise. The data warehouse absorbs data at such a rate that company officials can understand the impact of advertising campaigns (TV spots, online banners and so on) within 15 minutes of their appearance.
Data analytics also evaluates the effectiveness of those marketing programs. In one recent case, eHarmony mapped the cost of acquiring users via affiliates to conversion rates and lifetime value. The company determined that marketing approach brought low-quality traffic and decided to cut the affiliate program. The resulting $5 million in saving goes straight to EBITDA, Avedissian says.
Jonathan Goldberg, director of strategic customer relations at MicroStrategy, says eHarmony uses its technology to assess its marketing initiatives and improve the effectiveness of its matching algorithms.