Software as a Service Business Intelligence Continues to Draw Interest, Traction
A recent survey sponsored by on-demand BI provider Birst finds that nearly half of respondents would consider a BI solution delivered as a service. Top concerns remaining around SaaS BI are complexity of integration, security, and opposition from IT, the survey found.
- By Linda L. Briggs
- November 18, 2009
Nearly half of small and midsize businesses would consider a BI solution delivered as a service, according to a survey sponsored by on-demand BI provider Birst. Top concerns around SaaS BI, according to survey respondents, are complexity of integration, security, and opposition from IT staff, who often fear a loss of control over data.
In this interview, we talk with Birst CEO Brad Peters, who was senior director of analytics at Siebel Systems before joining the software-as-a-service provider, which was founded in 2004. Peters discusses some issues uncovered by the survey, including the misconception that BI rollouts need to take many months. That’s not necessarily so with a SaaS implementation, according to Peters, who says, “For the record, at Birst, the vast majority of our deployments are measured in weeks.”
The survey of IT and business professionals in North America was conducted by Toolbox.com. Nearly two-thirds of those surveyed represented small and medium businesses, and more than half of respondents reported using on-premise BI already.
BI This Week: The top three challenges in a SaaS implementation, according to the survey respondents, are complexity of integration, security, and IT opposition. Were you surprised that those three issues were topped the list?
Brad Peters: The top three challenges reflect valid concerns. Complexity of integration is a challenge for any BI deployment, whether it’s on-demand or of the traditional on-premise variety. Preparing data from multiple sources for analysis is a large implementation step, and companies may be curious as to how this step is going to be alleviated for on-demand users. Some solutions, like Birst, can easily integrate information from multiple sources, while others require that the customers do this on their own.
The second issue, security, is a classic hurdle for any SaaS vendor. Companies are entrusting critical data to you and want assurances that their data will be private and secure. Thankfully, it’s also an issue that is easily evaluated, since security requirements and protocols for SaaS are now well established, so this has become a single conversation, one which often follows a checklist. While it’s rightfully a major concern, it’s a concern that a company can get straightforward answers on fairly quickly.
IT opposition is a challenge that we’ve also seen alleviating in the past few years. Now that IT departments are facing budget and resource cuts, there is less of a desire to have direct control over all the technology a company uses. IT departments are increasingly embracing the concept of a solution as a service, since it means that they don’t have to buy and maintain hardware, manage software updates, and act as the front line of support. Now they can pay a SaaS provider to handle all of those things for them.
The survey showed that 45 percent of companies would entertain a SaaS solution. What’s holding back the other 55 percent, in your opinion?
Business intelligence delivered on-demand is still fairly young, and I think that companies want to see more proof of value before they get involved. Today, many companies would quickly consider CRM as a service, in the form of Salesforce.com. When Salesforce.com first started, they had to do intense sales and marketing efforts to get started and spread the word. We’re just coming up that curve with on-demand BI. It’s actually impressive that 45 percent of companies already would entertain a SaaS BI solution.
The historical experience with BI is that it’s expensive, complex, and time consuming to deploy and update, so there may be some reticence to try out a new BI vendor, even one with a dramatically improved value proposition. I think companies are still concerned about security and about how IT will be involved.
One survey surprise for me was that 70 percent of respondents expect that getting a SaaS solution deployed would take longer than three months. Where is that perception coming from? What is a realistic average time for a BI SaaS installation?
We were also surprised by that statistic. I think the market has been conditioned to expect BI deployments to take a long time. If you’re purchasing an on-premise solution, this is definitely still true -- getting a traditional system of reasonable complexity to deliver real value takes some time. While the market may already understand that SaaS BI is more affordable, there is some skepticism about deployment speed.
For the record, at Birst the vast majority of our deployments are measured in weeks, and are deployed well within 90 days.
More people are implementing, or have implemented, SaaS BI (16 percent) than in-memory analytics (4 percent) or open source BI (just 9 percent), according to respondents. What do you make of those numbers in terms of the strength of SaaS in the marketplace?
This was a surprising result for us. There are some very prominent in-memory analytics and open source BI companies out there, so we were expecting a much stronger showing for them.
I think the results show the compelling value proposition of SaaS. Open source BI has dramatically reduced upfront software costs, but the implementation and maintenance resource requirements are not much better than traditional BI. In-memory can be very affordable to initially deploy but doesn’t scale well and is not necessarily as powerful.
SaaS BI has power along with affordability, scalability, and ease of use. That message is spreading more quickly than we thought it would, perhaps because companies have a greater willingness to consider more options in order to achieve their technology goals.
Only 37 percent of respondents said they would use SaaS BI with a SaaS operational application such as Salesforce.com. That means 63 percent would use SaaS with an on-premise or in-house operational solution. Can you explain that response?
There was an assumption made early on about SaaS BI solutions: they were appropriate only for data coming from other SaaS systems, like Salesforce.com. However, this is simply not the case. SaaS BI can handle data from any system, either on-premise, as with an Oracle database or desktop Excel files, or from SaaS systems. This survey indicates that customers understand this – they know that BI solutions can pull data from multiple sources, and they expect SaaS BI to do that, too.
One concern that is sometimes voiced about SaaS is that IT can lose control of an off-premise solution. Did the survey address that?
The survey did address the issue of IT opposition, which is usually due to concerns about control.
It’s valuable to break that up that discussion into various components, however, since IT departments are increasingly willing to have the “plumbing” aspect of BI -- maintenance of hardware and servers -- offloaded to third parties. The remaining control concern is about the security of the data itself and ensuring that a SaaS system doesn’t somehow disrupt the company’s other systems.
With BI, this is less of a concern, since the security issues are now well known and addressed, while the source data is often kept in operational systems or other data warehouses from which the BI system is pulling. In short, BI is analyzing your data, but not interfering with it.
Over half of respondents said they would use SaaS BI in conjunction with an in-house BI/DW system. Is that a common trend among your customers? What size companies tend to use SaaS BI along with in-house systems, and in what ways?
This was an interesting statistic, since many Birst customers use our SaaS BI solution in conjunction with another in-house system. Historically, companies tried to build one single master BI system to rule them all, but quickly discovered that this was untenable due to cost and complexity. Now, there is much more of a portfolio approach.
Interestingly, customers are ahead of the industry analysts on this. We have many customers who say, “Well, I installed on-premise BI for the finance team years ago, but now I need something for marketing and sales, and it’s too expensive for me to extend what I already have. Since you’re compatible, faster to deploy, and more affordable, we’ll go with Birst.”
This is in contrast to questions we get from industry analysts all the time, who ask, “Why would anyone rip out their expensive on-premise system and replace it with Birst?” The answer is clear -- customers won’t rip out their old systems. They’ll just add Birst alongside the existing systems to address their growing list of BI projects.