Now let’s talk in terms of making a dashboard solution plug into predictive analytics. There is something called PMML. That’s industry standard language, it’s called Predictive Modeling Markup. And PMML is a way that you can exchange predictive models to run against standard databases. So a BI solution should support PMML.
Analytic solutions like SAS and SPSS can output PMML structures, and you just need to have a way to import them. There are a couple of issues there but delivery formats that are appropriate for organizations, so dashboards, alerts, mobile capabilities, time limits, we talked about.
There are lots of problems with spreadsheets. One of the things we see that distinguishes the innovative firms is they are delivering information in a much more timely fashion. You can see a huge difference between the innovative firms and the tactical ones.
Does timely mean day prior information? I would say close to real time is best. That means either daily or intraday. My own opinion is that close to real time would be several times a day. The data should be updated every couple of hours, or every x minutes or every x hours.
|#1 Ranking: Read how InetSoft was rated #1 for user adoption in G2's user survey-based index
This is not so much a software issue but a process issue. Once organizations have information available for them frequently, they realize a need to use it. Sometime we find that our customers who are using our dashboard product are expressing some level of frustration, they are not experiencing as much improvement as they hope to. But it’s because of the lack of timeliness of their data updates. If they can review the information more frequently and make it part of their organizational processes, they will improve the performance of the organization.
Making information more accessible, this is a huge issue. 89% rank that as important, make it easier, simpler to use a BI solution. We certainly say that our BI software is relatively easy to deploy and easy to use. That being said, a successful deployment needs to focus on best practices. And this is not best practices for our BI product per se, this is industry best practices. Data access continues to be an issue. People can't get to all the right data. That’s important. From a process, services standpoint, you want to make sure that different databases can be pulled together, or mashed up.
In large enterprises data is often stored in silos. There are ownership issues around the data. Sometimes they aren’t willing to share the information. This also varies by industries. So I know for instance, in the health care industry, there are often boundaries between the different data sources, due to regulatory issues. So you have to understand those are issues.
But the issue I am more concerned about as an organization is the parochial kind of siloing. For instance, in a manufacturing organization a department is not willing to share their quality control data because they don’t want people using it to criticize them. But you need that information, to understand various cost issues, and you need that information to understand some of the HR issues. If we are trying to figure out how many people to put on different parts of our manufacturing processes, we need to understand the quality control, trends and issues. Are they trending up? Are they trending down? That’s an impact on hiring.
But different parts of an organization will sort of hold their cards close to their vest. My favorite term for this, I heard somebody use this term once. It’s not just data segregation. It’s data balkanization. It’s like warring factions over the data. So this is related to the previous point where I said there’s a performance edge for using internal resources on BI projects versus relying on external resources.