March 17, 2015 | Written by: Kevin Allen
Share this post:
Twitter data isn’t just for marketers anymore.
This week, IBM and Twitter announced their latest round of cloud data services that will allow business pros and developers to gain valuable business insights from Twitter data.
Will Reilly, director of IBM’s business and Industry Solutions Marketing for Big Data & Analytics, joined us to talk about the new features, which will be available on Bluemix and Watson Analytics:
What should people know about the Twitter and IBM partnership?
That we’re making tremendous progress. We announced the partnership with Twitter in October of last year. This week, we’re announcing the availability of software services for developers and business professionals to help them gain insights from Twitter data. We also have over 100 client engagements in progress and we’ve trained over 4,000 IBMers to help those clients get value from Twitter data.
How are you seeing clients use Twitter data?
It’s interesting—one of the things we’ve found is that clients are moving from just using Twitter as a data source in customer service and marketing to broaden and deepen it—to start using Twitter data in other parts of the business. That’s something that IBM is uniquely positioned to help with, because to do that you need to understand how to manage the data, apply advanced analytics to it, and you need to understand business and industry processes to be able to apply the data back into other parts of the organization.
What are some examples of insights you’ve gleaned from this work so far?
The first one is about churn. This is something that can be applied in telecommunications companies and cable TV or satellite TV services—anything like that where you’re paying on a monthly subscription basis. All of those companies have churn models that predict, on an individual subscriber basis, your likelihood to move and to switch providers. What we found working with one provider was that we were seeing churn that we couldn’t explain. Two consumers with exactly the same data history and purchase history in the same marketing segment and the same propensity to churn, were churning differently. We were trying to figure out what was going on, so we started looking at external data—weather data, real estate data, Tweets—to see if we could understand what was happening. What we found was that there was a certain set of subscribers who were very sensitive to changes in weather and how weather was changing the service that they were getting from this company. And they were tweeting about it. We could then make a correlation between tweets and weather events and realized that that was a mark of churn that this company hadn’t identified before. Through that, we were able to help them improve their churn model, by around 5%
Any more examples you can share?
Another one is around the link between employee turnover in retail and customer loyalty. Retail as an industry deals with a lot of employee turnover—the nature of the work is that employees don’t tend to stay in those jobs very long. What this company wanted to figure out was whether high employee turnover impacted customer loyalty—and which customers does it impact the most? We found out that it does impact customer loyalty and customer satisfaction significantly. People tweet about the changing experience they’re having in the store. One of the things they write about is that the person they’re used to seeing very regularly—maybe on a daily basis—isn’t there anymore. Therefore, the service that they’re used to getting has changed, and they don’t like it. What we were able to do is realize that the people who are most impacted by this change are their most loyal customers—at least they were for this organization. So, what this company did was put in place some employee nurturing and training responsibility. It’s basically an HR story on how to retain your best talent.
Twitter can also be a pretty powerful demand signal. Could you share any stories where you’ve seen this play out?
We buy cars differently from how we used to. We do all our research online and we go to the dealer when we’re ready to buy a car. This gives them very little insight into their demand—what people actually want. Then customers tend to get upset when they go to the dealer and they don’t have the right vehicle, the right color or the right options. But people are Tweeting about the purchase process. So we’re experimenting with auto makers being able to use Twitter as an early indicator of demand that replaces how people used to shop—by going into dealerships and talking with reps about what they want.
With these new features, what will businesses be able to achieve now that they weren’t able to before?
One is we’re integrating Twitter data into some of our key cloud-based analytics software products. There are two services available now on Bluemix. Through those, developers can get access to Twitter data through a Hadoop environment or in a custom-created Twitter insights service. What that means is developers can start building data-centric applications that are built on Twitter data with all of the virtues of the Bluemix platform—speed, simplicity, scalability. We’re seeing clients already start to build applications on Bluemix efficiently with Twitter at the heart. One bank, for example, built a Twitter application in weeks rather than the months it would have taken them previously.
The second thing that people can do is for the business user. This week, we’re announcing the availability of Twitter on Watson Analytics, our cloud-based analytics platform for the business user. What that means is that business users can now use Twitter data in an easy-to-use analytics application.
The third thing people can now do is by integrating these tools with the work they’re doing with IBM Global Business Services, we can start to pull in Twitter data into all those interesting areas I was talking about, around HR, operations, supply chain and so on. To do that, you really need a partner with the know-how to be able to do that effectively.
Finally, what role does the cloud play in the Twitter/IBM partnership?
Cloud is the only environment in which we’re accessing Twitter data. You can only get access to that data through cloud-based analytics tools from our software portfolio. So, the cloud is really important in this partnership. It makes it so simple for clients, whether they’re developers or business users because it reduces the friction of getting access to this data set. It’s very large and unstructured, so it’s not the easiest data to deal with. We’re doing that work for you, basically. It’s now very quick and very easy to get access to Twitter data, and the cloud platforms are completely fundamental to that