Implementing a new business intelligence (BI) software is an exciting time. You start discovering all the data you’ve been creating, and everyone feels inspired by the possibilities that data can create. All of the BI software tools make it seem like you just download their software and you are immediately connected to insights. This is where you need to remember that these websites are written by marketers. Marketers love to show you the path to a solution and downplay the hard parts of implementation.
To help you better understand what your BI implementation process is going to look like, here’s some straight talk. We’ve also included an upside to all this, because every problem is also an opportunity for growth.
What to expect from your BI project:
- You will figure out that your data is messy
- You will be unclear about your goals
- You will have to limit the implementation at first
- You will need support
1. You will figure out that your data is messy
No matter how meticulous your company is about properly tagging and classifying customer data or ensuring that every single purchase is logged in the database, you will find that data, by its very nature, is messy. Missing fields, incomplete records, and duplicate entries can all muddy your data. Some of this is caused by human error: people aren’t great at filling in forms consistently, and sometimes your employees don’t have all the information necessary to complete all the fields. Degraded data—or bit rot—can also be caused by changing formats.
But you can clean up your messy data. Data cleansing tools can perform automated processes to help you improve the state of your data. At first, your data analysts may also spend a lot of time manually inspecting files for improvements, duplicate data, and incomplete records.
The upside to messy data
Data analysts are process improvement junkies. They want to make the data cleansing process easier on themselves, so ask them for notes on the most common problems they see. Then take these notes to the responsible departments to turn into policy. Behavioral improvements and a little bit of automation will ensure that the data you make going forward is cleaner and takes less of your analysts’ time to improve.
2. You will be unclear about your goals
Business intelligence software vendors (and their marketers) promise that you will gain “insight” and “clarity” from your data. That you will get instant forecasts and find those bottlenecks. I have written those exact words when talking about BI software. But what the marketers won’t tell you is that you have to know what you’re looking for before you find it.
Clear goals are SMART goals: Specific, Measurable, Achievable, Relevant, and Time-bound.
- We will improve the revenue from our core customer group by 5% in Q2.
- We will decrease the average turnaround on contracts by 15% in March.
The upside to unclear goals
There really isn’t an upside to unclear goals. But the improvements come when you realize your mistake and fix it. Hopefully, you will get halfway into your first project, figure out that your goals aren’t clear enough, and then you’ll start to understand how to clarify your goals in a way that uses your data to your advantage. This may take trial and error. Document what you’ve learned and use it moving forward.
3. You will have to limit the implementation at first
It’s tempting to purchase your BI software and immediately train the entire team on how to use it. But a widespread rollout is almost certainly poorly-planned and doomed to fail. If you’re really intent on building a data-driven business, you need to start small.
Build a single project for a single department that targets a single goal. This limits the amount of time your analysts spend on data cleansing and import, it focuses your team around building a repeatable process, and gives you the flexibility to react quickly to the inevitable unknown unknowns.
The upside to a limited implementation
By limiting your implementation to a single initial project, you reduce the stress on your analysts, build greater buy-in when the project succeeds, and waste fewer resources across the organization. It means a slower rollout, but these lessons will help your team scale faster in the future.
4. You will need support
You’ve hired the smartest analysts and purchased the best BI software available. The analysts can figure this out!
Sure, but each system is different. The power and ease of use of your BI solution will depend on lots of factors including:
- The querying language it uses
- Whether it’s a code-based, drag-and drop, or query-based interface
- Whether it has native data connectors or uses API connections
- Whether the tools are built for analysts or business users
You will need to consider the support options that the BI vendor provides, and whether your users have access to human customer support specialists to get nuanced answers to their questions.
The upside of support
Is there a downside to support? Maybe if there are limited support options in the BI software you choose. But the upside to getting good BI support is exactly what the marketers have been saying all along: faster time to insights, quicker setup, lower project costs, and more informed analysts. All of these lead to greater revenue.
Going from BI software purchase to practice
It will take time, but you’ll get there. A BI software purchase is an investment in growth, and it should be treated as the software project it really is. Start by outlining your goals, cleaning your data, and building out a single project with support, and you’ll be visualizing all the data in no time.
Need help finding the right BI software for your business goals? Our unbiased Technology Advisors understand your needs and will help you choose a short list of options that will get you analyzing your data faster. Fill out the business intelligence Product Selection Tool and a Technology Advisor will contact you with your short list.