Running a Proof of Concept: 12 points to consider when selecting a Business Intelligence software tool.
Proof of concepts (PoCs) are becoming very popular as organisations look to minimise their risk by trialling new business intelligence (BI) software before making a decision to buy or not.
The PoC does not need to be exhaustive (i.e. testing every aspect of the software), but should strike a balance between testing the technical features and measuring the business value, within a specified timeframe. Here are 12 introductory steps to consider when trialling a new BI software tool.
1. Limit your evaluations
Use existing market information (such as the Gartner Matrix) to help narrow your selection. Create your shortlist and if you evaluate two products you may want to stagger the trials.
2. Download trial versions
Make sure you have the relevant permissions to install the software on your computer. Check both administrator rights and system requirements before commencing the trial so as not to lose any valuable time.
3. Plan your evaluation carefully
Are you going to attempt to re-engineer an existing report or tackle a new challenge? Keep it simple and sketch out your process and success criteria beforehand; what would “good” look like? Rather than trying to rebuild existing reports, pick less complex, but important evaluation factors such as testing data connections, cleaning/blending data or building some new visualisations.
4. Engage key stakeholders
Make sure all parties (e.g. Finance and IT) are on-board with the proposal of a new BI tool, but do not involve them all in the trial as this will add complication and may hinder the process. Keep things simple by picking just one team to run the PoC, and ask each colleague to attempt a slightly different challenge. This way you will gain a broader (but manageable) view of the tool and its business benefits.
5. Data access
Do not worry about trying to bring all data sources together in one place. Modern tools can connect to the various data sources in situ, whether they be on your network, in the cloud, in a third party system and so on. Make sure you are familiar with the data you are connecting to beforehand, so you maximise the time evaluating the tool rather than trying to understand the data.
6. Clear your diary
Ensure you have allocated time for the trial – at least a couple of hours a day over the typical 14 day period. Consider working with the supplier/implementation partner to set a series of milestones to keep you on track. Make sure you also have time set aside before and after the PoC to allow for the necessary preparation and post-trial appraisal.
7. Train yourself first
The main dilemma in a PoC is how to deliver some meaningful outputs that give a fair reflection of the product’s capabilities when you do not know how to use the tools. Invest some time in watching training videos/reading blogs before the trial starts or make a small investment to have an experienced partner work alongside you.
8. Get support
Do not be afraid to ask for advice and support as you start to build out your analysis/reports. Waiting until the end of the trial could mean you miss out on many tips and tricks, and regular checks will ensure you are heading in the right direction.
9. Do not aim for perfection
Do not try and make the analysis or end reports perfect. Rather than producing a final deliverable that can be pushed into a live production environment, use this as a learning experience and then build out a full solution using best practice if the PoC is deemed successful.
10. Know your audience
Consider how your colleagues will be accessing the output. If the end deliverable is an interactive dashboard, think about the layout and usability e.g. laptop/mouse or iPad touch screen. Do not wait until the end of the trial to share your results. If possible, engage with colleagues and get feedback throughout to ensure a better end result.
11. Set clear criteria for success
How will success be measured? Is your ability to connect, blend, analyse and visualise your data good enough? How long did it take previously? Include ‘time spent’ as part of your business case calculation. Think about speed, ease of use, complexity of analysis, scalability and price.
12. Make your case
As part of your evaluation plan, build a business case to quantify the likely return on your investment. Can you make a case for ‘time to value’ of fewer than 90 days? How much time can be saved and how does this translate to a financial saving? Can you replace other redundant tools?
Following the trial, consider licence costs (make sure there are no hidden extras), your internal sign off process, different types of configurations (desktop vs server). Finally, ask for a demonstration of any features you did not get chance to test. If you have any doubts, make sure you check these with the supplier/partner rather than making an assumption that the tool can or cannot do something.
With more than 300 vendors providing data visualisation tools alone, the analytics and business intelligence space is very crowded. Follow these steps to ensure you choose the right tool for your business.