You’ve Decided on a BI Tool, Now What? 10 Steps to a Successful Roll-out (and Mistakes to Avoid) You’ve Decided on a BI Tool, Now What? 10 Steps to a Successful Roll-out (and Mistakes to Avoid)

You’ve Decided on a BI Tool, Now What? 10 Steps to a Successful Roll-out (and Mistakes to Avoid)

You’ve Decided on a BI Tool, Now What? 10 Steps to a Successful Roll-out (and Mistakes to Avoid)

Author: Eugene Khazin, Principal and Co-founder, Prime TSR

I see a repeating pattern across many companies today. IT decision-makers and C-suite executives get all fired up about BI tools and are onboard about implementation but have no game plan on how to make it all happen. 

  • Do they want to achieve better decision-making capabilities? Of course!
  • More efficient marketing and better transactional insights? Who wouldn’t?
  • Are they ready to write a check for a fancy new BI tool to turn their company data into actionable insights? Absolutely.

But at the end of the day, these leaders are unsure how to implement the BI strategy or tools that they purchase. They underestimate the complexity of implementing an enterprise-grade BI tool; the dependencies, functional requirements, questions, and other best practices that need to be carefully scoped out in advance. 

Gartner Research, in fact, points out that 70-80% of business intelligence initiatives end up failing. I see it happening every day. So, here’s my advice to the VP of IT or CxO decision-maker who has decided on a BI tool and is now deciding on what to include in the implementation steps.

The following 10 steps are my field-tested recommendations on what to include in order to ensure a successful roll-out of your BI tool strategy.

1. Start by measuring existing pain points

The first ingredient of success in BI tools adoption is the most obvious – but surprisingly the most overlooked: Identify your key stakeholders and understand what their pain points, needs, and expectations are. 

The decision has been made and the executive team is on board, but you need to work with the players on the front line who are going to be addressing complicated data problems, change management resistance, and user adoption issues. 

When determining your stakeholder’s needs, it’s important to consider the following: 

  1. What information do they need to perform their functions?
  2. What is their ideal future?
  3. What types of performance analysis must they achieve?
  4. What business units will be supporting each report? 
  5. How many reports will need to be developed and when?

You can start by measuring existing pain points across the company in order to identify a reasonable goal. And be specific. Here’s an example:

  • Pain point: Your team lacks clarity about customer behaviors on the corporate website
  • Solution: Develop targeted content based on real-time customer metrics that leads to a 10% increase in conversions during the 3rd quarter

2. Now, go identify how to solve those pains

Next, it’s important to clarify what steps the BI tool must perform in order to address the pain points and needs identified above. Some examples here would be: 

  1. The tool must have the ability to drill down, drill across, and slice-and-dice.
  2. The tool must provide Storyboarding, Geospatial Integration, Animations, and Barcodes. 
  3. It must alert business users according to user-defined parameters.
  4. It must allow full data integration in real-time from multiple disparate sources.

These functional requirements must also be backed up with specific technical requirements. For example:

  1. The tool must be available 99.99% of the time during any 24-hour period.
  2. The tool must include access to a variety of data sources including Big Data Connectors, Hadoop, Hive, MapReduce.
  3. The tool must support ELT (extract, load, transform) integrations.

3. Understand how the data will be stored

You need to clarify the core elements of your database and supporting architecture. Some key questions to consider here are:

  1. What data sources will you require (social media, emails, mobile data, etc.) 
  2. How will your data sources connect to the database? Will it be integrated via ETL or ELT or a hybrid? 
  3. What data marts and warehouses will you require (Hadoop, Cassandra, MapReduce)?
  4. What data element listings will you require (first name, last name, email, address) so you can find specific nodes across multiple systems?
  5. How many records will you be accessing?

4. Nail down your performance metrics

Choosing the right KPIs is going to be critical to the success of your BI strategy. Being crystal clear about your KPIs will help to rally your employees towards a reachable goal, rather than waste significant resources chasing after vanity metrics.

For example, if the company goal is to increase revenue by 20% over the next two quarters, you need to look at your average rate of return on existing customers or ask yourself whether more sales training makes a difference in attaining closed deals. 

Based on your specific corporate objectives, you may ask yourself the following questions:

  1. What metrics are we going to track? (web traffic, returning visitors, click-thru rate)
  2. What reports are we going to build? (analysis, charts, dashboards, drill-through)
  3. What are the measurements that give life to our metrics? (i.e., what were the marketing & sales cost for Q3) 

KPIs are the crux of your BI strategy; get these right and half of your headaches in implementation will be resolved. 

5. Plan for Data Accuracy

The volume, variety, and velocity of data today requires a rigorous data governance strategy. Stated slightly differently, any policy needs to present a unified and consistent view of information across the company. 

Some key questions here will be:

      1. What are your “sources of truth” for validating the data?
      2. How do we ensure our data is reliable, consistent, and repeatable?
      3. Who will be our stewards to ensure data integrity?
      4. How do we avoid GIGO (garbage in, garbage out)?

6. Avoid vanity metrics at all costs

Closely aligned with your data governance policy is the matter of your data origins. Can we trust the data, is it reliable and robust, and can it be validated? Company integrity is on the line, so it’s important to nail this from the start. Key questions are:

  1. What databases, software applications, and files will the data come from? 
  2. Who are the business owners and SMEs of that data?
  3. What will be the data integration and data quality requirements? 

Don’t be “that company” which mistakes vanity metrics for quality reporting; customers are too smart for that and will tell the difference. Reputations are on the line here.

7. Make sure your tool supports the business (not the other way around)

It’s great that you and your team have reached an internal agreement about the BI tool of choice. But ultimately you want to ensure that it supports your primary business objectives.

“Companies tend to focus on the user experience for their customers, but when it comes to internal apps for their employees, they tolerate clunky hard-to-use interfaces. User simplicity is also important for business tools—including BI because your employees are important users also” Praveen Puri, Puri Consulting

I’ve heard plenty of horror stories over the years from companies that went down this path and then found out too late the tool didn’t support a primary KPI. Be crystal clear upfront about what you need and expect on matters of functionality. Much of this is already nailed down during the business and functional requirements stage. But it’s always good to have a redundancy check.  

A case in point might be that you have an important KPI that includes a start and end date. But you find that the reporting tool vendor only offers the end date. When writing the SOW, be clear that you’ll need a customized solution to gather the key performance metrics that are going to be most important to your line of business. 

8. Identify and measure what success looks like

Your team has spent countless hours developing the BI strategy. Over a period of many months, they’ve hammered out the business objectives, pain points, functional specs, and KPIs. But at the end of the day, everyone has to agree on what success will look like for the company. 

Here are some categories to consider for measuring the success of your BI implementation:

  1. User adoption rates
  2. Financial performance increase
  3. Improved employee productivity
  4. A rise in customer retention

As you build out your “success measure” metrics, keep in mind there are three principles that I’ve learned to be always true in today’s data economy. Adhering to these will save you countless headaches:

  • Organizations implicitly trust Data 
  • Real-time Data is powerful 
  • Accurate Data is king

9. Ask the right questions

It goes without saying today that organizations that adopt BI tools just perform better on every level. And, of course, achieving operational efficiency is the goal of every company. But what does that look like on the ground and in the day to day? 

One of the benefits of BI integration is that business can finally ask the right questions. This is something that often gets overlooked in the rush to “always find the answer” to more revenue, customer retention, etc. 

The cause and effect of putting actionable insights to use can be powerful. Here’s a case in point from the standpoint of operations. A company uses BI to discover that everyone is clocking out at 6:30. So, it decides “why don’t we turn the heat down at 7 and save $1000s of dollars a year in heating expenses.”

Or, another case. Analytics data shows that the first stage of processing for insurance claims takes 8 days. Customer stakeholders start to ask, “can we bring that down to 2 so that we get a major boost in customer satisfaction?”

Always be looking ahead to the possibilities of what proper analytics reporting can look like for your company. Start by asking the right questions.

10. Make sure it will scale

When it comes to BI integrations, there is additional consideration that you’ll need to build into your SOW based on your particular company and industry. But the core ones that are very important across the board are: 

Security – Data privacy is very important all around but it is especially so in the healthcare and financial sectors. With the prominence and rise in enterprise hacking incidents in recent years, don’t leave the security of your BI tools or data to chance. 

Disaster recovery – What happens in the event of a major power outage or data breach? Data is changing in real-time all the time and as the volume and variety and velocity continue to grow, new disaster recovery scenarios must be envisioned.

Scalability & organizational roadmap – The simple consideration here is “will the tool grow” with your changing needs as a company? The picture in 12 months is going to be significantly different than today. Keep this factor in mind as you build out your SOW.

Now what?

It’s great that your organization has decided to build out a BI strategy, vetted the best solutions, and settled on a platform that will turn your actionable insights into gold. 

But this is where the hard work really begins. From here on out everything’s about figuring out the best practices, requirements, dependencies, and numerous other processes necessary to successfully get your BI tool successfully implemented. 

Unfortunately, this is also where the majority of businesses fail. 

But that doesn’t have to be your company’s story. In case you’re feeling stuck and unsure how to move forward with implementing your BI strategy, give us a call today. We can turn things around, help you scope out the SOW, and deliver a powerful implementation strategy that will meet your KPIs, ensure user adoption, and win over new customers in no time.