A few weeks ago in Redwood City, California, dozens of CFOs gathered for a Technology Forum seminar to explore strategies for financial transformation. At one point, a speaker asked the audience, “How many of you are worried about the quality of your data and what it means for your business?”
The room became a forest of hands.
It’s hardly a surprise. CFOs are worried about dirty data because of the problems it has long caused traditional financial planning and enterprise performance management (EPM) solutions. They worry because those solutions inevitably must integrate with other software, and wherever two applications meet, there’s data.
They may have been attending a forum on transformation, but these CFOs were feeling the strain of years spent under the yoke of EPM applications. Dirty data is terrifying to them because, in the past, software integration has been a struggle—a dark art that required a specially-trained guru who midwifed the integration through a long and painful birthing.
In the age of on-premise EPM and analytics platforms, the responsibility for integration typically fell to the customer. Again, not surprising, since these implementations were nearly always elaborately customized through months of hard-wired coding, making no two deployments alike.
Customers weren’t demanding that EPM vendors make integration easy—in that environment, easy integration just wasn’t realistic. So Oracle customers knitted together their applications using Fusion middleware, SAP customers called upon Netweaver, and IBM customers turned to Websphere.
See more: Is The Private Cloud Really Bigger Than The Public Cloud?
This had all developed before applications, services and data moved to the cloud. InformationWeek’s recently-completed State of Cloud Computing Survey found that of the respondents who use cloud services, 79% rely on multiple cloud providers. To get those cloud applications talking to each other, one-third of those organizations deploy connectors from vendors like SnapLogic, while 41% bravely take it upon themselves to bridge internal and external applications by custom-coding scripts and application stubs around vendor APIs.
I find this frustrating. The genius of the cloud is that it provides an opportunity to transcend the earthbound limits of on-premise platforms. It shouldn’t be up to cloud customers to figure out how to get their cloud solutions working together. Yet, here are cloud customers, still carrying the burden of integration. It’s as if every new television owner also had to purchase a soldering kit and hard-wire the components inside to get it to work with their cable box or DVR.
Yes, data must be exchanged between applications. And the truth is, there’s no recognized standard for cloud app integration. But cloud solution providers can make integration easier—even transparent—for customers. So much so, in fact, that software integration isn’t something today’s enterprises should even be talking about.
In other words, integration may have been the customer’s headache back in the days of on-premise solutions. But in the cloud, it’s up to vendors to solve it.
Unification, not integration
Unlike legacy platforms, cloud applications make it possible to unify rather than integrate different software solutions. Unification requires an entirely new approach, a complete rethinking of how data flows from one application to the next. And it requires hiding the details from customers, creating a unified user experience that conforms to the way people want to work, rather than imposing clumsy new processes, widgets or other extra steps that stand in the way of productive use.
So what should CFOs look for in a cloud-based EPM solution? What are the markers of cloud applications built for unification, not integration? Three crucial attributes come to mind for truly transformative EPM platforms, and I believe they should serve as a prescriptive for all cloud applications moving forward.
1. No more data mapping
For decades, data and software integration has been defined by the familiar ETL (extract, transform, load) process. But ETL is inherently limited because the transform step applies a series of hard-coded rules or functions that makes the data useful within a target application.
This data mapping can take months to complete, and once it’s done, it’s static. So unless you’ve predicted with startling accuracy every possible way you’ll need to use that data in the future, you’ll inevitably have to rebuild that data map to make it current. And unless you’re sure your cloud applications will never change or update, you’ll have to rebuild that data map. And unless your business landscape remains unchanged for years to come….well, you get the idea.
See more: The Hand That Rocks The Hybrid Cloud Could Rule The Enterprise
Data mapping is not a process that moves at the speed of business. An alternative approach is to extract your files and data and then load them into a secure, cloud-based storage platform like BOX. Then, when the application has to execute the data, you can transform it on demand. ELT (extract, load, transform) offers a more fluid and flexible approach, letting you load as much data as you want and then transform it in ways that meet your needs as they arise – no mapping required.
2. Built for collaboration
A McKinsey Global Institute analysis of IDC data found workers spend 18% of their work week communicating and collaborating internally (it jumps to 46% if you include communicating via email). So any transformative EPM solution must have collaboration in its DNA.
This is different from simply posting a spreadsheet in the cloud so a handful of stakeholders can view it. This is about having the built-in ability to work in real-time with unstructured data so multiple parties can update, correct and improve the same document at the same time. It’s about having an application that accelerates business, rather than getting in the way of it.
3. Industrial strength, without industrial weight
The advantages and flexibility of the cloud doesn’t absolve EPM vendors from having to support complex transactions. EPM applications still must stand up to the most rigorous, destructive, incremental and bulk data load tests. They still need the no-compromise reliability and integrity of legacy EPM. Fortunately, when you shift from classic integration to cloud-based unification, you can retain that reliability while gaining cost efficiencies that were never possible in an on-premise paradigm.
See more: Google Ups The Ante Against Amazon’s Cloud, And That’s A Good Thing For Business
Premise-based EPM implementations must be designed with high levels of redundancy to ensure they are always available in case of power failures or other problems, so customers are saddled with an overbuilt solution that costs a lot to build and maintain. Cloud-based EPM applications, in contrast, are designed to anticipate and instantly respond to the failure of a cloud node. So a catastrophic event in one data center won’t bring your EPM application down.
When EPM vendors deliver cloud applications without forcing customers to focus on integration, they can let CFOs focus on what really matters. And before long, questions about data quality barely register a shrug. And that’s exactly as it should be.