What was the scope of their involvement?
We’ve taken a project-driven approach with Keyrus. They work on some of the more strategic initiatives we’re working on. The goal is to have them engaged on projects where we need focus; we don’t want them flip-flopping between too many competing priorities, so we can have a dedicated team.
Specifically, the projects they’ve been working on with us have been in two buckets: One is around machine learning, so they’re helping us deploy machine learning algorithms. We’re using their resources to build decision trees built in Python, do data cleansing and learning, and to generate outputs from various input data sources. They’re required to produce some accuracy measurements to give us a sense of confidence as to how good the machine is operating. We share those metrics all the way up to our executive level.
The other stream is on our power BI deployment. We’re moving away from a traditional BI model, and away from our Oracle BI reporting platform. We’re moving into deploying power BI as a more modern, visual tool. Keyrus’ expertise in that modern space has helped us a lot.
They come with strong developer resources, as well as strong project management resources to deploy reporting. They also help optimize our data models, and we want to continue to accelerate that. We’re just building upon our partnership and upon the knowledge that they come to the table with; they know the beverage and alcohol space fairly well. For that reason, they’re able to accelerate a little quicker, when it comes to deploying new data models on that platform.
In the past, working with other vendors not keen on our industry, we experienced a longer data education process. Keyrus, however, helps us accelerate the process and we use them to deploy what we call “self-service BI,” in terms of training and enabling our business to build their own reports. With this process, the business doesn’t have to always come to IT to build new reports.
What is the team composition?
There are four key people on our team. The first two are a project manager and a lead developer, who have the largest time commitment.
We also have an account executive, who serves as a main point of contact if we have an escalation; she also joins our weekly calls and any upcoming work is typically discussed with her.
Lastly, we have an SME. That person is someone is that’s familiar with the alcohol and beverage industry. They’re also the lead of the data science practice areas. Not only do they assist me in contextualizing the nature of the data, but they also come in when we need their machine learning expertise.
We also have a very specific data scientist person that we’re not engaging right now because that part of the project is closed, for the time being. They do very specific Python programming and all of the data cleansing for the machine learning engine.
How did you come to work with Keyrus?
We heard of them through an affiliate. We’re a global company, and one of our other branches used Keyrus in their region. They put their North American representation in touch with us, and we went with them. At the time, we were unhappy and in-between BI partners; we shifted to them without too much of a competitor search.
How much have you invested with them?
We’ve invested over $500,000 with them.
The first year was around $150,000. In the last year, we used a lot more budget because we have those two initiatives that have ramped up recently.
What is the status of this engagement?
We began working together in June 2017, and our engagement is ongoing.