Okay, so you have tons of data and are in need of capitalizing on big data and analytics. Currently, you likely are manually populating Excel spreadsheets or have a BI tool and are trying to determine the best path forward.
Whether you are just getting your feet wet or knee deep into it, we’ve developed a list of five basic building blocks to successfully move analytics forward at your organization. By ensuring each item on the list is checked you will be well positioned to start reaping the benefits of your market’s recent explosion of data.
First and foremost, securing buy-in from high above is a must. To initiate a successful endeavor into analytics, key decision makers in your company need a clear understanding of the value of analytics as well as the process by which they are best created. If the leaders in your company cannot fully understand their value or view analytics as just a bunch of pretty pie charts, then you have a tough road ahead.
Ideally, buy-in comes directly from the CEO. At the very least, gain the support of an executive who has influence and the ear of other key executives. Quality analytics aren’t created overnight. It is critical to team up with leadership who appreciate the potential value and understand the time needed to develop the proper foundation.
Given the speculative nature of implementing advanced analytics, visionary leaders can support the beginning groundwork by helping tear down internal barriers, and actively taking advantage of the insights produced. Be sure to keep your executive support engaged throughout the process, as they can provide invaluable feedback and guidance to ensure that full analytic potential is realized.
Okay great, you made it past step one, have secured executive buy-in and, are ready to begin … but where to start? Even with all of the advances in data capture and the dramatic improvements in hardware/software, the old adage still applies: Garbage In -> Garbage Out.
You likely have plenty of data. Your concerns now are: Do you have good data? Is it easily accessible? Timely? Are you capturing the right elements? Is it storing historically or is it simply appearing and disappearing just as fast?
Often, the answers to these questions aren’t black and white, but more a shade of grey. The existing data is good for answering one business question but not others, or it has the capacity to answer a critical business question but not in the timeframe needed for a material impact.
Data availability and quality is a journey, not a destination. I’ve worked on large teams with large budgets and there was always something with the data we wanted to improve, capture, or enhance. Regardless of where you start, take one data point after another and always be on the lookout for ways to improve.
I can’t emphasize enough this crucial building block. No amount of high priced software or legions of highly educated Data Scientists can overcome rotten data. However, if you find yourself in this position, don’t despair, you are not doomed. You will just need to spend more time getting the data to a place where it is usable before you dive too much further.
Over my many years of analytics experience, I have had the pleasure (and sometimes the pain) of working with a diverse list of analytic software vendors. The list of features these softwares have can be impressive, but it’s also just as important to understand what they can’t do.
Having a clear vision of where you want to end up is critical. Talk to executives, end users, consumers, etc. to find out how they prefer to interact with the information. Do they prefer paper reports or are they more likely to use a mobile app? Are they Excel junkies who want to download the data and blaze their own path or are they wanting the analytics to do all the work? There is a wide range of possible deliverables so understanding your organizations preferences will be key when evaluating software vendors.
Once you have a better sense for where you want to end up, start narrowing down the list of toolsets to consider. Firsthand knowledge from co-workers and peers can provide great insights, as well as research companies that produce reports comparing and contrasting similar vendors.
Personally, I have honed in on a few key toolsets that are tried and true and have allowed me to meet and exceed very demanding environments. TARGIT Decision Suite, for example, puts straightforward self-service analytics into the hands of decision makers, empowering them take action on their data.
When evaluating your situation, you will most likely need several different toolsets to get the job done. For example, you may need tools for storing data (RDBMS), moving and transforming data (ETL), visualizing data, performing advanced analytics, etc.
Make sure you have a clear understanding of each tool you are evaluating, strengths and weaknesses, and what they can and cannot do. Many vendors offer free trials which I highly recommend taking advantage of. Also, don’t be afraid to look into open source software as not every company has a big budget for analytics and open source tools like MySQL and R can offer great returns for their price point.
People are critical. Never underestimate the benefits of having high caliber, intelligent, and innovative people to work with. Having worked as a one-person startup to the other extreme of directing an international team of almost 100 people, the success of any analytics project will rise and fall in direct correlation with the quality of people assigned to it.
Don’t be seduced by where they went to college, a fancy title like Data Scientist, or that they match your toolsets exactly. Look for and hire people that are intelligent, ask inquisitive questions, can give specifics about actual experience solving analytic issues, and—perhaps most importantly—are a good fit for the culture of your team and organization. Nothing can ruin team dynamics more than a bad fit for the team no matter how well intentioned.
Another consideration is speed of implementation. The size of team needed to build the solution may look much different than the team needed to support it. Using consultants to augment your team during development can help ensure timelines are met without adding long-term costs.
Perhaps this last checklist item should be a sub bullet to first checklist of Executive Buy-In but I think it is important enough to have as its own entry. Doing analytics right isn’t cheap. Some of the best tools available can easily run six figures and hiring and retaining talented staff can be especially costly given their high demand and low supply existing in today’s economy.
However, no matter the size of your budget, there are budget-friendly options out there. Open source software offers robust functionality at a cost that can’t be beat. Other more expensive tools are shifting their business model toward subscription-based pricing which makes it more affordable to get started by paying a small monthly fee versus a giant upfront cost.
In terms of staffing, focus on finding the right mix of skillsets and personalities … not simply on hiring an army of people with advanced degrees. A small team that works well together, compliments each other’s talents, and is given the freedom by management to explore and innovate can achieve amazing results.
The last question you may be asking is can I get by with four out of five items checked? As with most things in life, there are always exceptions but with each unchecked item the path becomes significantly harder to navigate.
For example, if you have poor data quality that is missing key elements no amount of money on tools or people are going to be able to overcome that. Likewise, you may have great data and high-end tools but if you don’t have quality people manning the wheel it is going to be near impossible to keep the ship moving in the right direction.
One last piece of advice I have worked for organizations that have had all five of these items checked and we were able to do some truly magical things there that I am still proud of to this day. Conversely, I’ve also worked for organizations where I’ve been told these items exist in the interview process only to find out it was all lip service. Needless to say, analytics was doomed from the start when it became clear they had very few of the checklist items needed.
I firmly believe in this checklist and it has allowed me to grow and develop professionally and I hope it benefits you the same.