Proof of Concepts (POC) designed to evaluate Business Intelligence solutions are a great way to mitigate the risks, understand the capabilities and identify the limitations of BI solutions.
Although POCs have different interpretations, depending on the business area and technology, a BI POC is used to validate a proposed solution through a demonstration with a small number of users.
There are various reasons why POCs are needed for BI implementation in an organization. Some focus on the end users, while others may deal with the ETL process or ad-hoc analyses. Some BI POCs may be small & quick, done within a few day, while others may span a few months and get quite iterative. Some may be initiated informally and executed as an ad-hoc project while others may require strict processes and funding comparable to a formal engagement.
Here are 5 steps to make the most out of your BI POCs:
1. Focus on value, and less on mechanics
It’s essential not to lose sight of the big picture, no matter how simple your BI requirements, or how informal the process. Many times, organizations focus on technical details before considering the overall business objectives to be accomplished with your BI project. For example, your BI solution may allow you to change the font size on your reports, but does your team really care about it?
A BI tool is only a means to accomplish your business goals that will grow your organization. If it can’t help you do it, then it doesn’t matter how many bells & whistles are on offer. The key is to start from the final business objectives and work backwards. This will help you avoid all the irrelevant features while executing your POC and quickly reach a conclusion.
2. Identify all the BI layers involved
Every BI project involves multiple layers – data, ETL, reporting, ad hoc analyses and more. Depending on the size and scope of your project, it’s important to identify which layers of your BI project need validation during your POC. For example, you may be evaluating only report design capabilities in your POC, but if your BI solution doesn’t support your data sources, then your BI project is bound to fail, unless you put additional ETL processes in place to transform your data before feeding it into your solution.
3. Cheat on sample data but not on business logic
POCs are limited in scope and need to be executed on an aggressive timeline. So you may need to cut corners while creating sample data to run your POC. That’s perfectly fine. However, don’t cheat on your business rules and concepts. For example, if your reports need to use a formula to calculate a metric, don’t hard code the final values instead. You may not be able to calculate it in your actual BI project because of missing data points, something that should have been caught in your POC.
4. Clearly define the scope of your POC
This is one of the most common reasons why BI projects fail. No matter how small or informal your POC, ensure that you clearly define and maintain your POC scope. This will help you identify all the required data inputs, transformations and reporting capabilities that need to be supported by the proposed BI solution.
This will also make it easy to communicate your requirements to the BI vendor and get an accurate cost estimate for your project. Otherwise, you’ll discover new requirements during actual BI implementation which will not only delay your project but also inflate your budget.
5. Staff the right talent
Identifying the right talent with the right background is critical for the success of your BI POC, as well as your BI implementation. At the minimum, your POC team needs to have a combination of subject matter experts in BI as well as those with in-depth experience about business processes.
They should also have good communication skills, required to translate technical details into business requirements and vice versa. It’s great if they also have good people skills, as they’ll need to interact with various stakeholders (senior management, BI solution providers, data vendors, business teams), and reach a middle ground whenever conflicts arise.
A POC is only as good as its assumptions and the team that executes it. So if your aim is to enable faster, better decisions, then technical knowledge alone won’t ensure successful implementation. It must begin with a business leadership that can pave the way to execute a BI vision where technology is a channel to deliver business growth and profitability.