As the adoption of Big Data tools accelerates, many of the success stories we hear are the big wins, the victories that Hollywood movies are made of. Think of Billy Beane and his use of data analytics with the low-budget Oakland A’s.
For most businesses, though, swinging for the fences all the time may lead to a pile of strikeouts. Instead, many businesses would be better served by focusing on small wins that they can build on. After all, before you can consistently hit home runs, you need to be able to get the ball out of the infield.
If you can make your supply chain a little bit more efficient, or you can streamline hiring, or you’re able to find a small change that lessens the abandonment rate of online shopping carts, these wins can accumulate into something much bigger.
In the mid-1980s, organizational theorist Karl Weick found that when people take on huge social issues – ending homelessness, eliminating poverty, stopping climate change – the very scope of the challenge often impedes progress. Weick argued that these world changers should focus on more manageable goals.
+ ALSO ON NETWORK WORLD: 10 big data startups to watch +
More recently, psychologists Teresa Amabile and Steven J. Kramer found that steady, ongoing progress is the key to employee morale. If you feel like you’re making progress, however small, at your job, you’ll be more satisfied. They call this the Progress Principle.
In other words, incremental progress is more important than some grand vision. So, if your goal is a lofty one with no clear roadmap attached to it, like “let’s dominate the server virtualization market,” well, you’re probably not going to.
Rather, you’d be better off targeting one small thing that you know you can do better than your rival. After you accomplish that, you gain confidence, generate good will, and tackle the next thing.
Here are examples from Airbnb, Starbucks, Sonic and Double Dutch of five small wins your company can achieve and then use as building blocks to gain momentum and make those lofty goals more reachable.
1: A picture is worth 1,000 words
Many startups believe they’ll be able to use Big Data to knock off major incumbents like Cisco or Google or Apple. They believe they’ll peer into their Big Data crystal balls and discover secrets that will magically catapult them into big-time winners.
That rarely, if ever, happens. Rather, Big Data success stories typically start with small questions – what’s the best city block for a new retail location; how do we turn the choosing of locations into a systematic process; how do we make it more likely that our sales teams will reach people when they call them; how do we change our retail offerings, in real time, to better match consumer preferences?
Look at one of the biggest sharing economy successes, Airbnb. It took the company a good long time to build up momentum, but they did so by discovering the one key obstacle that was preventing people from using Airbnb rather than hotels.
Riley Newman, head of analytics and data science at Airbnb, ran a regression to determine the features of a listing that had the most impact on whether it would get booked, and he unearthed a shockingly simple insight: unappealing looking listings don’t get booked. Appealing ones do.
That’s it.
So, Airbnb started experimenting by sending professional photographers to various properties to redo the photographs, and the results were off the charts.
Airbnb eventually rolled out a tool that allowed hosts to request free professional photo shoots, and in the process, not only did they boost bookings, but they earned a ton of loyalty from the owners.
With that win under their belt, and with plenty of insight into what works and what doesn’t when presenting listings, Airbnb has continued to lean on analytics, which now drive everything from its search algorithms to marketing efforts to customer support.
2: Finding the perfect location
We’ve all heard that real estate mantra of location, location, location, but a good location is often deemed so only in hindsight. Our second Big Data lesson comes from a retail giant that figured out a process for better determining where to place new locations.
In the Internet age, many businesses focus too much on digital strategies, while ignoring the fact that e-commerce still only accounts for 17% of total U.S. retail sales. In other words, people still do the majority of their business in meatspace, not online.
When brick-and-mortar businesses seek to expand in order to capture a bigger piece of that multi-trillion dollar pie, a major challenge is figuring out where exactly to locate a new store. In the past, most prospective business owners would target an area that looked like a good bet. They measured street traffic, or pedestrians per hour, or they simply looked at what other businesses looked like in the area. And, of course, the basic availability of office space was always a driving factor.
However, just because a site looks like it will be a good location doesn’t mean it will be. That similar business that looks to be thriving could actually be losing money, or if you target an area that appears to be underserved by music venues, perhaps it’s because music venues don’t play well in this area.
Often, miscalculating by even a few blocks could mean the difference between success and failure.
This is one reason why Starbucks relies on data analysis to guide the process, going so far as to build a market-planning and store-development application called Atlas. The best way to think of Atlas is as a Big Data analytics tool layered on top of mapping software. Built on Esri’s ArcGIS (GIS stands for Geographic Information System), Atlas lets Starbucks factor in a range of variables that contribute to the success of existing stores, visualize them on maps, and then seek out similar areas for new locations.
At last year’s Esri User Conference, Patrick O’Hagan, Manager of Strategy for Starbucks, walked through a typical scenario. He fired up Atlas and pulled up a map of Nanning, China, a city of 2 million people where Starbucks already has eight locations. O’Hagan ticked off various “layers” that can be visualized on the map, each of which influences store location. These include proximity of trade areas, demographic information, daily traffic volume, and the availability of public transportation.
EsriO’Hagan zoomed into a district of Nanning where three new office towers would be opening within the next two months, representing a promising potential location.
Once a new store site has been targeted, a workflow screen pops up, which guides the process of approving a site with corporate, securing the proper permits, and then actually launching the store.
But Starbucks doesn’t stop there. Here in the U.S., faced with a saturated market for coffee shops, Starbucks is using Atlas to help it roll out new menu items, such as offering beer and wine sales at certain locations.
Pulling up a map of the Los Angeles area, Laurence Norton, BI Director of Development and Engagement for Starbucks, illustrated the variables that factor into site selection for this pilot program.
“This map shows pilot locations along with wine-away-from-home purchase patterns. As we look to roll out the Starbucks evening menu to more and more locations, we can target existing coffee houses in areas with high spending patterns,” Norton said.
Leaning on data and maps won’t guarantee that any of these efforts will succeed, but the process greatly reduces the risks associated with a new store launch, stacking the deck in Starbucks favor.
3: Power dialing for prospects
Double Dutch, which provides mobile social networking applications for conference and events, was being pushed by its investors to generate more sales. The startup had been slowly but steadily growing, but for its investor base, the slow-and-steady approach wasn’t good enough.
DoubleDutch had raised $6 million from venture capitalists, and needed to prove to investors that they would indeed deliver a return on their investment, explained Russ Hearl, vice president of Worldwide Sales Development. “The best way for me to provide a return on their investment was to build an outbound sales machine.”
So, in early 2013, DoubleDutch began building a new sales development team to help accelerate revenue. Like many high-growth technology companies, it decided to use sales development reps to contact and qualify its leads before passing them to its closers.
The trouble is that you actually have to have conversations with those leads before you can convert them, and the irony is that in this over-connected age, decision makers are increasingly difficult to reach.
DoubleDutch sought out a solution that would help its sales development reps connect with more prospects in less time. It figured that if they could do this one small thing – have more conversations with prospects – they’d inevitably be able to close more deals more quickly.
So, DoubleDutch hunted for a service that would combine automation with predictive analytics. They ended up using PowerDialer from InsideSales.com, an app available through Salesforce AppExchange.
PowerDialer lets sales teams dynamically sort targeted contacts, pinpointing not only hot prospects but also the reachable ones. PowerDialer relies on machine learning to identify sales patterns by analyzing millions of anonymized profiles and sales interactions.
Sometimes, the sales tactics it unearths seem like common sense – although when no one else is exercising common sense, can you really call it that?
For instance, one factor that boosts reachability is a local number. If you’re a sales rep calling from an 800 number or even from a number with an unfamiliar area code, people are less likely to answer. They don’t know who you are and probably suspect that this is a sales call.
A local number, in contrast, boosts reachability, since people assume they may know the person on the other end of the line, even if they don’t recognize the number. With PowerDialer’s LocalPresence feature, when a sales rep calls a prospect in a different area code, a local phone number will be displayed on that person’s caller ID.
This and other features – including automated dialing, automated text messages, detailed reports, and even gamification features – helped increase DoubleDutch’s dial-to-connect rates by more than 60 percent, which led to an increase in conversions of 37 percent. Overall revenues also increased by 200 percent.
“We did a test where we had some of our sales development reps using the PowerDialer and some not,” Hearl said. “We found that the ones who were using the PowerDialer were booking about four times more demos, and they were generating about four times more pipeline.”
4: Tweaking the menu
We’ve all seen this movie before: a venue you love faces a tough economic climate, so they start cutting corners and raising prices, and in the process, destroy what was great about the place in the first place.
The fast food chain Sonic did not want to fall into this trap. With more than 3,544 drive-ins in 43 states, Sonic has turned the nostalgic drive-in concept with carhop delivery service into a competitive advantage.
Sonic competes in the extremely crowded Quick Service Restaurant (QSR) segment, where margins are often razor thin. To keep its edge, Sonic realized it needed to find new ways to absorb rising costs without passing them onto customers.
While Sonic could see the potential its vast data stores offered, it was frustrated by the performance limitations of the analytical solutions that it was using. Sonic wanted quicker access to high-value insights, but it was reluctant to invest in a complex and expensive solution requiring multiple vendors and long development cycles.
After rejecting a number of Hadoop-based solutions, as well as ones that would require mixing and matching various vendors, Sonic settled on the Big Data analytics engine from 1010data, a tool that is capable of processing billions of rows of data almost instantly. Because the 1010data analytics engine wasn’t built on an aging relational database model, it didn’t require workarounds to compensate for the limitations of legacy architectures.
With the 1010data analytics platform, Sonic can perform queries on all of its data, not just subsets that had to be prepared beforehand. Sonic was also able to gather insights more effectively because 1010data’s user interface provides a spreadsheet-style view of the data that simplified queries and made the data accessible to nontechnical users.
“1010data allows our analysts to interact instantly with all our data. By giving them the freedom to explore and improvise, the platform encourages discovery, and makes data analytics a factor in every single business decision,” said Craig Miller, Sonic’s senior vice president & CIO.
+ 10 Top Big Data Stories of 2014 +
Support for “ad hoc,” on-the-fly queries allowed Sonic’s analysts to quickly model programs and menus that strike a much better balance between ROI and customer value. Rather than alienating customers through rising costs, Sonic was able to tailor menus to what customers wanted, while finding other ways to protect their margins.
5: Sharing insights via the cloud
For Sonic, the initial success of their Big Data program inspired them to seek out other areas where they could put analytics to good use. Sonic is now considering extending the 1010data analytics tool to its franchise community.