News Stay informed about the latest enterprise technology news and product updates.

Adoption of self-service BI tools requires convincing

Getting business users to move away from spreadsheets toward dashboards can be challenging, but pays off in the end when departments collaborate more.

LAS VEGAS -- For Heather Campbell, implementing self-service business intelligence tools at the fundraising office...

at Princeton University wasn't the hard part. The real work was actually getting people to use the system.

Campbell -- director of analytics and data management at the development office at Princeton -- spoke at Tableau Conference 2015 last week, telling attendees that many of her business users were so used to looking at spreadsheets that they didn't want to use anything else. But doing data analysis in spreadsheets is laborious, as most columns are likely irrelevant to users, which makes it hard to quickly see outliers or trends.

Starting around 2012, Campbell's team broadened the office's use of Tableau and built some dashboards that allowed users to track the performance of giving campaigns, or model the effect of having scholarship recipients send thank-you notes to donors. Still, adoption of the dashboards remained slow.

Use of the self-service BI tools has since picked up, but there was no one silver bullet that spurred adoption. Instead, it took a combination of training and collaborating with business users to ensure Campbell's team was delivering relevant and useful dashboards. Over time, confidence in the dashboards grew.

"Data strategy is a driving force at our organization, and it now feels like we've reached a critical mass," she said. "We accomplished this with one focused dashboard at a time."

Aiming for day-to-day use of Tableau

Implementing self-service BI tools is only a small portion of the battle in becoming a data-driven organization. The real work often comes in getting workers to actually use the software. Users at the Tableau conference shared their tips for getting frontline employees to embrace the tool and adopt it in their day-to-day work.

We accomplished this with one focused dashboard at a time.
Heather Campbelldirector of analytics and data management at the development office, Princeton University

Jason Flittner, senior business intelligence engineer at Netflix, said self-service implementations work better as a collaboration between developers and users. Gone are the days of traditional BI, in which the development team gathered requirements from users, spent weeks or months building a report and then delivered a final product. Flittner said these days, developers need to sit with business users and take feedback in incremental steps to continuously work toward something useful.

For example, he recently built a dashboard to allow a business analyst to view data on what kind of Netflix shows people are watching, how often they skip ahead during certain programs and what shows people tend not to finish viewing. All of this information tells the business about what kind of shows people like and want more of.

The initial dashboard Flittner delivered was, for the most part, static. The analyst was happy with it, but not thrilled. In response, Flittner made a single change that allows the analyst to filter the data by each row, which was a simple revision. This change made the analyst much happier with the product.

"This is what has helped us get our analysts more comfortable with visualizations," Flittner said. "It's really powerful."

Seek self-service BI champions within departments

Kate Treadwell, a consultant with the firm InterWorks in Stillwater, Okla., recommended using extensive training after implementation to drive engagement. This doesn't just mean a quick how-to for developers. She said people from all departments should be involved. Folks from marketing or sales might not need to know about all the functions of self-service analytics tools, but allowing those people to see what the software can do might get them excited and encourage them to reach out to developers with requests for dashboards that they'll actually use.

The other benefit of getting people from all departments engaged, Treadwell said, is that you find "champions," people who like the idea of self-service analytics and encourage their teammates to also use the tool. There's only so much an IT department or dashboard developer can do to get individual lines of business to use a tool, she said. The people who work in those departments are going to be more influential, and convincing just one champion about the merits of a dashboard can make a big impact.

"They're your voice when you're not around," Treadwell said. "Having someone on your side is one of the most important parts of an implementation."

Ed Burns is site editor of SearchBusinessAnalytics. Email him at and follow him on Twitter: @EdBurnsTT.

Next Steps

Everything you need to know about self-service BI

Self-service BI requires flexibility and strong governance

Keep IT involved, even when moving toward self-service

Dig Deeper on Business intelligence software

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

What are your tips for spurring adoption of self-service BI tools?
The first thing is, BI vendors have spent 25+ years pursuing the "ease of use" to get more people to use BI. What they missed is what I call ease of usefulness. What can you do with the tool? With legacy BI vendors, ot's pretty prescribed, but the so called data discovery tools, like Tableau, have opened that up. Access to more data, especially via DIY data prep tools, is driving more DIY BI. But it isn't quite that simple. I have a white paper about this on Slideshare:
I am looking at exactly this  for my thesis. It seems that there are at least two things going on here, the first is the well documented cognitive benefits associated with visual analytics (expressed in the article in the oft heard complaint around the limitations of tables in excel). There is other more subtle but equally important epistemic work being done in the course of the development of the dashboards and that is contained in the notion of collaboration. Here the dashboards are acting as boundary objects between two communities--in the article it is between the developers and the business users. In my research I am seeing some dashboards being used across several communities of practice or , more accurately, functional groups, and these are helping build a common understanding which facilitates co-ordination with only a partial or no understanding of another practice. To understand this dynamic I am having to delve into organisational knowledge creation and diffusion theory and I would be very interested to hear if others had similar thoughts. I am particularly interested in how a dashboard embeds novelty (say in an unexpected insight derived from the analysis of the data) into it and thereby creating and diffusing new knowledge... what impact does this have on organisational practice? So in effect the dashboards are OD interventions    
The suggestions in the article are not terribly different from how we implemented (successful) BI programs. I find again and again people characterize BI programs as "IT-centric" and taking months to get a report. That hasn't been my experience in over 20 years with BI and it feels like the rhetorical technique of Mal adsurdum, the mischaracterization or invalid reduction to an absurdity. Characterize what has gone before as what it's not to make your point. 
I'm sorry, but I have a real problem when the word "force" is used. You should never have to use it to convince a user to adopt something. It never works. If you can't convince them, then it is very likely it was a "pork" project that was not needed in the first place. You get push back, and you get company politics involved in something that should have neither one. If you don't have buy in from the user community first, then they won't support you nor use what you just built them without a great deal of resentment.