Novartis uses Snowflake to bring life-saving medicines to market quicker

Mark Samuels Profile picture for user Mark Samuels April 4, 2022 Audio mode
Summary:
As well as a boost to data-led research processes, the Snowflake platform is helping to improve Novartis’ internal business operations.

Medic using healthcare technology on tablet © metamorworks - shutterstock
(© metamorworks - shutterstock)

Life sciences giant Novartis is using Snowflake technology as part of an integrated approach to cloud and data that is helping the company accelerate the development of medicines.

From initial research through manufacturing, trials, and distribution, it can take around 12 years to bring a new drug to market. By applying data and artificial intelligence to these processes, Novartis believes it can reduce the time to nine years. Snowflake is playing a critical role in this potentially life-changing process, says Loïc Giraud, global head of digital platform and product delivery at Novartis. He says:

The biggest benefit that you get with Snowflake is rapid access to information, whether that’s data ingestion or data access. Ultimately, when the users use a product, they just want to have a seamless experience. They don’t want to wait for 10 seconds, 15 seconds; they want to click on a button and get the result. That’s what the Snowflake platform provides and helps our engineers to integrate our information in a manner that we could not do five years back.

Back then, Novartis was largely reliant on on-premises infrastructure for its data analytics and data science processes. This approach hindered responsiveness, says Giraud, creating too big a gap between the demands of the business and the speed at which teams could conduct tests and bring new products to market. 

The company decided to take a fresh, cloud-based approach and adopted Snowflake in 2017 as part of an ongoing and broader effort – known as Formula One – to digitize every aspect of the company’s operations and bring data to the heart of its processes:

We used to have a big data platform that was based on Hadoop and the key challenges we had were performance and scalability. The initial reason we started to look at Snowflake was to get a scalable system that was high performing.

Giraud says the first process his team investigated was compensation cycles for field representatives. The previous system the company used relied on a cumbersome process that required a significant amount of precision. That first use case proved the benefits of Snowflake to senior stakeholders in the business and sponsored a wider implementation. Giraud explains:

We were able to deliver so much value from the ability to run through cycle times much faster than we had in the previous system. Then, over time, we built more and more use cases. Now, across all the areas where we operate, Snowflake has a portion of the service that we run on the cloud.

Now, Giraud says Snowflake is ingrained into Novartis’ business processes. The technology is used in three ways – for analytics generation across multi-cloud platforms including AWS and Azure, to ingest data and prepare it for insight generation, and to help the company share data across departments and out across a wider health ecosystem. He adds:

The goal is to try to look at different data providers across public and private offerings, which contribute value to our organization. It just makes it simple. And we also share our data with others.

Giraud says Snowflake has been integrated seamlessly into the company’s cloud-based approach to data ingestion and sharing. As with any new technology implementation, there was a “ramping-up period”, where people using the platform had to learn fresh skills. However, these challenges were met head-on, and the increased use of Snowflake has been pushed by the business rather than the technology department. Giraud explains:

I think the value you get for the performance is what attracted all our users. In fact, our users asked us to use Snowflake before IT had even looked at it. We did one or two use cases, and it was relatively smooth. The product is very stable, which is not always common when you work with a new company. You can scale up and down in an almost unlimited manner.

Eliminating non-value added work

Giraud says Snowflake is now an integral element of Novartis’ best-of-breed approach to its IT architecture. For each of the technology capabilities the company requires, the business analyzes whether a native cloud provider can provide the required features. Snowflake is used in the business’ refinement layer, which is about creating aggregated data to serve analytics generation, insight preparation and data-sharing processes. As well as playing a crucial role in reducing the time it takes to produce drugs, Snowflake is also helping to refine and improve performance across business operations, says Giraud:

Take finance – we used to do a bottom-up finance forecast. You have a lot of controllers and the only thing they do is crunch data day in, day out to build up your monthly forecast. Now, this forecast is completely running on AI – the controller just does adjustments; they are not doing the data-crunching. We’ve eliminated a lot of non-added value work by just automating certain processes, curating our data set, and making it transparent and available for users to consume through Formula One, of which Snowflake is a key component.

Further developments are ongoing. Novartis recently joined Snowflake’s newly-launched Healthcare Data Cloud, which aims to create a single, integrated, and cross-cloud data platform to eliminate technical and institutional silos. Giraud says this new development will provide another welcome boost to his company’s data-sharing activities:

Instead of creating a bespoke solution, and then being the first one to create something or the second one or even the third, we can work with others to develop industry standards that can be used by others. So, that's why we believe this Healthcare Data Cloud is important.

Five years on from the initial deployment of Snowflake, Giraud says the platform is helping the business to put data at the core of its business processes. He advises other digital leaders who are thinking of exploring the technology to also adopt a use case-driven approach. He says:

I think it's important that you don't just implement the tech; you're implementing a scenario that delivers a business outcome. And for that, you use your technology to deliver this outcome. Make sure you deliver incremental business benefits. Don't spend a lot of time putting things together before the business users start to see something. And then use an Agile and product management-led approach.

Loading
A grey colored placeholder image