Since its founding by two pharmacists in 1945, Alcon Labs has become a global leader in eye care. Now serving patients in 140 countries, Alcon Labs is at the cutting edge of research into new and innovative eye treatments, including their famous LASIK surgery.
Robby Dally, Director of Global Insights and Analytics for Alcon, recently spoke with us about technology’s impact on their business and the importance of making great use of data and AI.
Dally, a self-described “data wrangler,” is a huge fan of data and analytics. He likes being on the forefront of automation, democratizing AI usage throughout the organization, and coming up with new AI use cases. “I love analytics,” he said. “We want to make decisions based on data, not on gut feelings or what we’re hearing from the eye surgeons we talk with.”
But Dally is not the usual kind of data scientist. He’s an autodidact who earned his data spurs building SQL databases many years ago. He admits to a few gaps in his software expertise. “That’s where a tool like DataRobot liberates me,” he says. “I don’t have the bandwidth to go learn Python, so I can use DataRobot instead. I use Alteryx to get my datasets right, then run them through DataRobot and let DataRobot do its magic.”
Alcon Labs has two core areas of business. One is focused on eye surgery, such as tools to help surgeons perform LASIK surgery or cataract repair. The other is focused on producing medications to correct problems, such as dry eye, and on the manufacture and distribution of contact lenses.
Managing a complex supply chain of high-tech products isn’t easy. To help with sales forecasting, managing finances, and gaining a better understanding of industry trends, Alcon Labs relies on the DataRobot platform.
Making sense of trend lines is one example. It’s one thing to add data to an Excel spreadsheet and model it. But to contend with a wide range of variables, Alcon also develops time-series models that consist of data points ordered by time. These models are designed to help forecast events influenced by factors such as time of day, seasonality, or autocorrelated target variables.
“In our case, different products can have different patterns, with different seasonality or different orders of magnitude.” Dally told me. “We sell a lot of capital equipment to smaller companies, and those sales can be pretty infrequent. Forecasting – that is a challenge we’re trying to solve.
“Or if I have a new product launching, I have zero past observations to help me make a future prediction. How do you handle that? We’ve been lucky to get a lot of input from the DataRobot team. We just keep iterating through things, and the more you learn, the better you get at this.”
Like every other company on the planet, Alcon has seen its business disrupted by COVID-19. Some of the products Alcon sells are used in cataract or other eye surgeries that, while important to the patient, are elective procedures. “In some places, elective surgery was just shut down,” Dally says. “So, all the patterns we had were destroyed. The flip side is this is a procedural business. People don’t stop having cataracts. At some point, they’ll come back.”
For now, Dally’s team is taking a hybrid approach: running models that include pre-COVID and post-COVID data and attempting to use that data to create an AI model that takes recovery into account. “Every month we refresh the models to see what happens,” he says. “That’s where we are right now.”
When we spoke, Dally told us that Alcon was running some 600 projects on our platform. Even so, choosing which AI models to pursue takes some thought. “We could have a cool idea, but it takes ten years, so no point to that,” he says. “Learning how to frame ideas has been really important.”
And ideas don’t come only from data scientists. Alcon was an early adopter of machine learning, data science, and AI and many of their employees are steeped in technology. As a result, great ideas often come from people who know Alcon’s business challenges, but are not data or AI experts. “Inspiration can come from a lot of places,” Dally says.
What does the future hold for Alcon Labs? Dally hopes that within five to ten years, AI becomes a broad, powerful tool for managing complex forecasts and processes. His hope is that AI can take over many of today’s mundane tasks, freeing people to focus on projects that have more value. “I’m not sure what that future looks like,” he says. “But it’s going to change the company and many companies around the world.”