October 31, 2016
By Sharath Gokula, Director of Analytics, Analytic Mix
“If you can’t explain it simply, you don’t understand it well enough” – Albert Einstein.
In the same vein, strong analytics vendors should be able to explain the complex statistical and technological terms to their clients in a simple manner. In fact, they should be able to clearly speak to the business value their analytic solutions provide. Analytics technology is a means to an end, and not an end in itself. At the end of the day, your analytics solution should be able to deliver an increase in ROI, and not merely add to your technology stack.
In turn, as business folks, you need to ask the right questions to understand how your analytic partners will add value to your business. This post gives an understanding of the nature of the questions to ask your analytics vendors. Your questions should center around the data, output, methodology, and differentiation of your vendor. Let’s dig deeper into each of these categories.
Where’s the data coming from?
Although we live in the world of advanced statistical procedures, the quality of the input data strongly determines the quality of the output. You need to first understand the data that is going into your analytic solution. You need to understand how much data is needed, as well as how granular it needs to be. Will the solution give you strong insights with a month’s worth of data, or does your provider need multiple years of data? Do you need data for every ad or will daily or hourly data suffice? You also need to understand the source of the data to determine the quality of the data. To operationalize your solution, you can ask about how your pre-existing system will integrate with the vendor’s system.
How actionable are their insights?
Next, you need to understand the output from your vendor. What insights and business decisions can you make based on the insights? Are these the important questions your organization is dealing with? Additionally, you need clarity around the reliability of the insights. Will these findings hold for the future?
To understand this, you need to inquire about the data validation strategy being employed. Finally, you need to be able to understand the visualizations. Do they convey the story? What is the technology for storing, refreshing and sharing the visualizations? How customizable are the visualizations? How scalable is the technology behind the visualizations?
What’s their methodology?
Further, you need to understand the methodology. If it’s a black box solution will the inner workings be explained? Will you get a clear picture of the assumptions made in the algorithm? You also need to find whether the solution is customizable and that it mirrors your internal operations.
What makes them different from their competitors?
Finally, you need to understand how your prospective vendors are different from the competition. There are often differences in the data collection procedure, methodology or business model. From a business model standpoint, analytic vendors use either a consultative approach, a self-service or a hybrid approach. You need to understand what suits your business needs. Most importantly, evaluate the team you will be working with – how much experience they have, and their past accomplishments. Asking for previous deployments will validate your vendor’s credentials. Don’t be afraid to ask for an instance when an analytic solution was not successful.
When evaluating vendors, it’s easy to miss the forest for the trees. So, remember to use all these pointers while evaluating your vendor and do not lose focus on the business value the analytics solution will provide.
Do you have questions for us? We’re happy to answer them! Contact us today.
About the Author:
Sharath Gokula, Director of Analytics at Analytic Mix, is a technology-driven marketing analytics expert with more than a decade of experience in delivering client-centric business intelligence.