grandeduc - Fotolia
As companies begin moving more of their IT infrastructure to the cloud, they run up against the challenge of efficiently integrating these cloud applications to their existing infrastructure. Many organizations are starting to look at integrated platform as a service (iPaaS) technologies to meet customer relationship management needs.
These kinds of tools are a natural fit for the cloud since they plug into many cloud services to begin with. Leading iPaaS products include Informatica Cloud, Dell Boomi, SnapLogic, MuleSoft CloudHub and Jitterbit.
"It is good to consider iPaaS, but you would like to do so with your eyes wide open to your overarching integration needs across the organization, so you can see opportunities to get more leverage and economies of scale from those tools," said Benoit J. Lheureux, research vice president, responsible for the application infrastructure and middleware practice at Gartner. "If you don't at least consider that, you have the risk of creating line of business or application development groups adopting different technologies and creating a mess over time."
Starting with CRM needs in mind
Mozy, an online backup service, turned to Dell Boomi when they ran into challenges keeping the CRM system up to date with new customers. Mozy Corp. has about six million customers ranging from consumers to large enterprises that store about 90 petabytes of data in Mozy's cloud service.
Mozy's old PHP-based infrastructure was not able to scale efficiently to keep pace with new customer acquisitions. Sometimes it would lag six hours behind and up to several days at times. The problem was the design was not multi-threaded, which made it difficult to add new servers.
"We had to look at different integration options and figure out which was the best fit," said Mark Goetz, manager of eServices at Mozy. "With Boomi, we did not have to take care of hardware, software upgrades or maintenance, so we could focus on making our business process work."
Turning to an iPaaS allowed the company to get visibility into the number of transactions processed, the health of the overall interface and the failure rate. Of particular importance was the need for prebuilt adapters, and centralized logging. The iPaaS also needed to support different types of connectors for consuming Web services, using file transfer, database connectors and a Salesforce connector. Mozy also needed a hybrid offering since some of their infrastructure was on-premise. "We needed it to work seamlessly across cloud and an on-premises infrastructure," said Goetz.
The CRM integration was just a starting point for Mozy, and Goetz' team members gradually increased their Boomi footprint as new integrations were required. They wanted to limit the amount of custom coding they were doing. "In many cases, Boomi is now our default integration platform because the management of integrations, monitoring and reporting is centralized, and it provides an easy-to-use interface," said Goetz.
Salesforce does come with a number of integration options, which makes sense in a few cases. But in many cases, Goetz has found that Boomi offers better security and is easier to manage and integrate. "If there are concerns around security or complexity, then we default to Boomi," he said.
Boomi also comes with data mapping recommendations, which make it easier to map data between applications for case objects and lead objects. Goetz said, "We found that those mappings are extremely accurate. In more than 90% of the cases, we are just taking over those mappings. This reduces the time we need to invest in researching the mappings and is a confidence boost for the team."
Boomi also helps Mozy's team resolve error messages. Goetz explained, "Quite often, the error messages you receive are common across Boomi customers, so Boomi can make recommendations about how to resolve common error scenarios. This is a significant productivity improvement."
Integrating for predictive analytics
The retail giant Bonobos Inc. was looking for a way to build a new data mining platform from scratch to leverage data from a wide variety of cloud applications. The company uses AWS for its website, CommuniGate Pronto for email, Salesforce for customer service, Simparel for ERP and a few dozen other services in the cloud.
"Building a traditional analytics platform that leverages so many disparate sources did not seem feasible and would have been outrageously expensive," said David Glueck Sr., director of data science and engineering at Bonobos. "The strategy that made the most sense was to use an integration tool in the cloud since all of the services were in the cloud already."
SnapLogic allowed Bonobos to build all of these integrations with a team of only four people. They chose SnapLogic because it had prebuilt connectors for its shared services. "This was a huge timesaver for getting the platform off the ground and also adding new sources," said Glueck.
He added, "As a developer, you spend part of the time developing integrations and part of the time developing utilities to make it move faster. We wanted a tool where you hit the ground running, and SnapLogic does a good job at that."
He also liked the intuitive mappings that made it easy to automate the creation of new integrations. "If you are working with a tool where you have to do those individually, it ends up adding a lot of time and effort," said Glueck. He made extensive use of SnapLogic SmartLink for connecting from one application to another. He found the tool was able to look at naming conventions and could guess which ones connect to another.
They are currently modeling about 1 to 2 terabytes of data, but Glueck expects this to grow to between 100 terabytes and 1 petabyte over the next year as they start to leverage more granular data from customer interactions. A big part of this will be for better understanding customer behavior as they glean better insights about their interactions on the Web and in response to highly targeted email campaigns.
For example, they hope to better understand if a customer abandons a shopping cart because Bonobos did not have the right size, did not like the product or were unhappy with the webpage. These insights will also help Bonobos management to determine which products they should be buying and how much.
The initial improvement from the iPaaS came from better automation so that IT could spend more time getting insights from data analysis rather than gathering data. Over time, Glueck expects this will lead to better sales and engagement.
Glueck recommends that companies looking at iPaaS tools should build their integration strategy with the idea of change in mind so that the integration platform is flexible and extensible. He said, "A mistake I have made in the past is to build a technology integration platform, but then it becomes rigid. The reality is that with today's business, the sources will change, and usually the technology teams are behind the curve. You want to build an architecture that is flexible and [makes] it is easy to replace a certain database or technology stack."