So you know you need business intelligence software.
You need to get more value out of the data you collect. So you’ve vetted your list of vendors, you’ve been through the proof of concepts, and you’ve landed on a product that’s right for your company.
The next stage, however, can either make or break your purchase: BI implementation. Implementation is where BI’s potential is actualized.
If the rollout and implementation fail, then that investment you worked so hard on might become part of the roughly 70% of BI purchases that fall through.
Business intelligence software licenses average about $1,500 per user, per year, so even a handful of users represents a chunk of money for a small or midsize business (SMB). Five yearly licenses, if not used correctly, can mean $7,500 disappears into the ether.
Check out these tips for what you need for a successful BI implementation so that it will be the start of something good, rather than the end of a painful process.
Tip 1: Know what you need
In last week’s blog, I argued that anyone shopping for business intelligence software should follow Socrates’ advice and “know thyself.” The same advice holds true for implementation.
For your BI implementation to be successful, you need to know what “successful” looks like for your business.
Elena Rowell of Looker suggests that this starts early: “Even before you define what key performance initiatives (KPIs) you want to track, the first thing is to define internally what success looks like for your BI tool.”
“This will help create a set of shared expectations for the product, streamlining the implementation process. Be sure to define those expectations, as well as a few of the project’s deliverables: for instance, “every department will have two to three key dashboards, or users will (or won’t) have the ability to do ad-hoc analysis,” Rowell continues. “Many times, defining what a product won’t do is just as important as defining what it will do.
Another key to BI implementation is knowing which specific factors, or KPIs, you need to focus on to improve your business, argues JP Lessard of Miles Technologies. That might be churn rate, conversions, or your profit and loss margin.
“Figure out what factors you need to look at, and say, if I can answer these questions, I can make better decisions, ” Lessard says.
The answers to those questions are the KPIs you want to track.
Bob Pepalis of Izenda corroborates the need for SMBs to know what they want to track before they begin BI implementation. “The most painless business intelligence software implementation experiences we’ve shared with our clients have been when they come prepared.”
An SMB is prepared when they “know what questions they’re trying to answer, and just why they need analytics with their applications,” says Pepalis.
As an example of this preparation, Pepalis cites one company that wanted to change analytics solutions. The potential client needed a program that could provide real-time data. They also needed “to manage resources—people and equipment,” for several public safety agencies. Lastly, they needed “a solution that could bring together the data from those different agencies that have their own applications and databases.”
Because the client knew they wanted to track those specific things, the implementation was easy.
Pat Hennel of Silvon also agrees. “Having all the necessary requirements from stakeholders up front saves a lot of extra rework down the road.”
And it’s the involvement of those stakeholders that’s also vital to BI implementation.
Hint 2: Get everyone on board
In the same way you need a definite knowledge of the KPIs you want to track, you also need a definite knowledge of who your advocates are. If you’re not the owner of the SMB, make sure the owner, and other leaders, will advocate for the implementation.
Like Ranjit Satyanath suggests at Computer Weekly, getting executive sponsorship is a key way to ensure success.
Corey Mendelsohn of NewIntelligence also recommends using “senior level execs as champions to level” your BI implementation.
Embrace collaboration as readily as these perky stock photo models. They’re just lousy with collaboration!
BI vendor Matillion suggests you look for advocates among the “domain experts who understand the reporting requirements in the business.” In many cases, these experts are the people who had been “hand-cranking management reports using Excel.”
As such, these advocates will likely know what data the business has, and how to get value out of it. These advocates can begin as BI missionaries, and as the implementation rolls on, help others learn to use the new software.
Another way to ensure new users will stay users is to make sure implementation, and use, are as simple as possible.
Hint 3: Keep it simple
It doesn’t matter how brilliant something is: if it isn’t easy to understand, it’s not going to get popular support.
I’m dating myself with this reference, but this is the “Fantasia” principle: you may have a thoughtful, well-designed product, but if it’s so complicated the customer can’t understand what’s going on, the product will be ignored. Disney’s “Fantasia” is an engaging, beautiful movie, but audiences didn’t know what to make of it. The result? Financial flopburger.
Artistically influential, sure, but notice that it took them several decades before they made a sequel.
There are multiple ways to simplify your BI implementation process.
First, make sure you buy a program that’s easy to use. Features like a drag-and-drop, visual interface, such as Alteryx’s, or natural language processing search, like AnswerRocket’s, will encourage hesitant employees to use the program.
In fact, ease-of-use is the number one criterion that bigger enterprises look for when shopping for BI software, according to the folks at Gartner (content available only to Gartner clients, but it is well worth it).
Gartner analyst James Laurence Richardson suggests that you make ease-of-use a priority, as “problems encountered with the ease of use of BI&A (business intelligence and analytics) platforms are the most detrimental to business benefit gained.”
Matillion suggests that choosing a program with “an easy-to-use front end, devoid of technical terms, table names, and SQL syntax, is critical to ensuring a successful business intelligence implementation.”
Once you’ve selected the right program, make the experience of learning to use the program as simple as possible.
One means of simplification Matillion recommends is to provide a “glossary of fields, measures and terms using the familiar nomenclature that the user already understands.”
Lessard seconds this idea, and notes that “a lot of BI software programs will let you set up your KPIs, and define them” within the program.
With that glossary available, an otherwise alien program will feel more like a part of business as usual, rather than something completely new and different.
Lessard also suggests a basic orientation to the software: “You don’t have to do in-depth training, but you do have to orient your users, and get people familiar and comfortable with the program. If they’re not, they’re not going to play around with it.”
Similarly, Rowell suggests other solutions, from ongoing “office hours” offered by more advanced data analysts, to online tutorials, to weekly training sessions with analysts. “Data can be intimidating,” she says, “and you don’t want people to stop trying.”
View learning to use business intelligence software as similar to learning a subject in high school: if you struggled with a tough subject, you probably went to extra, after-school help. The same strategy will help you in implementing your software.
It’s that willingness to play and experiment that sets up employees to make the sort of ad hoc queries at the heart of self-service BI.
Hint 4: Start with quick wins
Your BI implementation shouldn’t tackle the entire company at the start.
Instead, roll out the program to one department, and log some quick wins. Those wins will be the best argument to get other employees to use the software. The old cliche is true: nothing succeeds like success, especially when you’re trying to move your business from their old ways to a data-driven, BI-empowered approach.
“Rather than trying to force the whole company to make a complete paradigm shift at once, rolling out the BI application to individual departments… allows you to focus your resources a little at a time,” argues Hennel.
Rowell adds, “quick wins tend to help BI spread organically—people see it and get excited, and want it, too.”
Rowell notes that this is how Looker often sees their software spread in successful implementations.
In the same way implementation should start with one department, it should start with one set of data. “Start with what you’ve got the most info about—that’s your sandbox,” says JP Lessard.
Look to turn that large store of data into a quick win before you expand the program to the rest of the company.
SMB owners have “a gut feeling” about what data’s important, and it’s that important data they should start with. “Learn how to use BI with that data, and that way, when you move onto other data you’re not as used to, you’ll be more ready to use the tool,” Lessard advises.
Gartner analyst Neil Chandler makes a similar recommendation (content available to Gartner subscribers), saying it’s a good idea to “conduct at least one new BI and analytics ‘quick win’ initiative to validate your analytics strategy” during the first 100 days of a BI program’s implementation.
Hint 5: Make sure the data’s clean
If the data you import into the program isn’t clean, you won’t get the results you’re looking for. And that can cost your implementation a lot of employee support. “The old garbage-in-garbage-out principle is so important,” Lessard argues.
If the data that goes into your BI program is dirty and inconsistent, the focus of implementation will instead be on cleaning that data.
Will the “Keep Calm and” trend ever end? Probably not.
Fortunately, ensuring clean data doesn’t require you know how to code. Though the phrase sounds technical, clean data simply means your information is entered in a way the BI program will be able to understand.
For example? “People will put text in a date field in their CRM software, rather than the actual date. They’ll write ‘next Tuesday,’ and the BI tool will ignore it,” Lessard says.
Other forms of “dirty” data are data entry mistakes (forgetting the period in a dollar amount, for example), or inconsistencies, like similar values with different names (e.g., one spreadsheet that labels employee pay as “payment,” where another calls it “paycheck” or “wages”).
And if the BI tool ignores it, that data won’t be properly uploaded. Clean data, in a lot of cases, means proper data entry. “Making sure your salespeople are using their CRM in a consistent manner is way easier than trying to clean the data later in the BI tool,” Lessard says.
Hint 6: Implementation is ongoing
Software implementation’s a lot like Christmas: stressful, demanding, and time consuming, but ultimately worth it when you see the looks on people’s faces.
Also like Christmas, the actual event is technically a one-time thing, but the spirit behind it can (and should) last all year. In the case of the Christmas spirit, this means finding ways to be generous year round, like letting someone into your lane in traffic, or not employing a creative hand gesture when someone doesn’t let you into theirs.
In the case of software implementation, it means being agile, and open to change. In the same way the actual event of implementation requires you to be organized and willing to do things in a new, data-driven way, that same spirit can (and should) influence your day-to-day work.
One of the benefits of an SMB’s agility is that bouncing an idea off of someone doesn’t require request forms in triplicate. So, if there’s a metric you think your business should be tracking, take Billy Joel’s advice and “tell [your supervisor] about it.”
When you don’t have to cut through a bureaucracy to get an answer, you’re already a step ahead of the enterprises.
Agility is especially important for an SMB, says Elena Rowell. “With the way a small business’s data constantly changes, the metrics that are important to you today aren’t necessarily the same ones that will be important to you in six months.”
As such, most SMBs find themselves in situations where agility isn’t just beneficial, it’s necessary.
Ultimately, business intelligence isn’t as much a set of tools or products you buy as it is a mindset, and a willingness to continually improve.
What are your tips for successful BI implementation?
Is your business intelligence software program up and running? If so, are there any tips you have for potential buyers? Let other readers know in the comments below!
Interested in other parts of the business intelligence buying process? Check out one of these posts:
Looking for Business Intelligence software? Check out Capterra's list of the best Business Intelligence software solutions.