For Abhinandan Bhattacharjee, unlocking the value of data is more than a business tactic, it has the potential to enhance lives. As a data engineer at UK-based online medical services provider Zava, Bhattacharjee sees this impact first hand. The company deploys a highly scalable, asynchronous model to analyze online survey data and match patients with doctors, who then diagnose and prescribe. The model allows for a single doctor to provide up to 1000 consultations in a single day, without the need for face-to-face meetings. Zava also leverages data to discover new markets and customers, especially those that are not aware of online medical services.

“Zava's main mission is to provide patients with healthcare at a fraction of the cost,” explains Bhattacharjee. “This adds meaning to my work because it's not just looking at data points, it's about helping people in their health, and that's fulfilling.”

For years the company warehoused their data in Amazon Redshift, and relied on managed ETL provider Alooma to clean, transform, and process data from the databases and event streams. After Google acquired Alooma in 2019 they dropped support for Amazon Redshift in favor of the BigQuery ecosystem, and Zava saw their data future in peril. With just three months advance notice Zava was left blindsided and scrambling for a new solution.

“It was a dark and scary time,” recalls Bhattacharjee. “Alooma was working so well for us that we considered shifting to BigQuery, but after realizing Alooma's future lacked a roadmap we were wary. We also felt secure with AWS, there hadn’t been any hiccups with the service, and a lot of our divisions and teams used interconnected AWS products.”

Zava's search for ETL alternative becomes an opportunity to rethink data

In order to keep their pipelines running Zava had a short time frame to migrate from Alooma, and their next ETL provider had to offer two “must-have” features: an advanced transformation layer that would allow Zava to easily parse, clean, and structure their data, and pipeline operation capabilities like error handling and pipeline orchestration.

“During our search we found many providers that offered one or the other, but never both in the same product,” explains Bhattacharjee.

An online query led them to Etleap, the one solution on the market that offered Zava's two must-haves, and whose similarity to Alooma gave them hope for a smooth transition. Like Alooma, Etleap is a managed SaaS offering with advanced transformation capabilities and the ability to monitor and maintain pipelines automatically. Both Alooma and Etleap are cloud-native, hosted solutions and support many of the same sources. They both aim to make ETL pipelines maintenance-free and easier to set up while still granting transparency and control to users. Zava decided to test the product for a trial period, and quickly realized the benefits of Etleap's ETL solution.

“Three weeks into the trial and we knew Etleap had everything we wanted and even improved upon Alooma,” says Bhattacharjee. “We were sold.”

Intuitive “Data Wrangler” offers ease and flexibility

The standout feature for Bhattacharjee and his team is Etleap's browser-based transformation tool called the “Data Wrangler”. Etleap's wrangler-based approach to ETL allows users to see a sample of their data and preview transformations in real-time. Then, users can define and map to their destinations (i.e. “wrangling’) without writing any code. Even non-engineers, such as data analysts, can construct purposeful data pipelines in minutes. This is especially useful for dealing with semi-structured data, where the schema needs to be teased out through parsing and structuring. And since users can preview how the script will affect their data in real-time, the Data Wrangler is a great place to reason about required data transformations and handle transformation errors as they arise.

“We love the DataWrangler,” says Bhattacharjee. “It’s great at handling JSON and transformations, and supports Python and custom functions, which is really good for our engineers, as we didn’t have to rewrite our Alooma functions from scratch.”

Etleap’s approach encourages iterative pipelines development where transformation scripts can be easily modified in response to source schema changes and data parsing errors. The ability to write functions in Python (and four other supported coding languages) benefitted Zava because they could simply import functions they had relied upon in Alooma, making slight format modifications and achieving the same result. This coding flexibility allows Zava to address unique business use cases.

Robust and secure ETL enables patient-centric care

After gaining a deeper understanding of Etleap's capabilities Zava has identified areas where data can be mobilized in ways they hadn't previously considered. Etleap's ability to pipe data from multiple sources into multiple destinations is a definite improvement upon Alooma , and is being considered as a method to enhance the surveys patients fill out to match them with optimal healthcare solutions. The surveys are the core of Zava’s customer experience, but presently the survey data enters into AWS Redshift in a semi-structured state. Because of Etleap's agility in handling semi-structured data, Bhattacharjee's team is exploring the benefits of bringing the data directly into Etleap. Once in Etleap the data can be used for predictive modeling, helping patients minimize the time it takes to fill in the survey, while at the same time providing more accurate diagnoses.

“Etleap is robust,” explains Bhattacharjee. “Its agility with semi-structured data and JSON allows for a much more patient-centric approach. We want our questionnaires to be shorter, and understand user behavior as they are taking the survey and feed that data back in real-time. Previously, we were not able to narrow in on all the complexities.”

Additional Etleap features such as support for S3 data lake destinations and the Data Modeling tool unlock the potential to draw further insights from their data streams. And because Zava manages sensitive patient information, Bhattacharjee appreciates that with Etleap patient data from their website will remain secure, confidential, and anonymous.

“Etleap is very secure and doesn't store the data in their system,” he explains. “Rather it stays within our network and the ETL tool is there to help us clean it up and make it ready for analysis.”

Vital pipeline prioritization and unparalleled support facilitate smooth migration

During the migration process Etleap worked with Zava's data team to identify sources needing support, existing Alooma pipeline configurations, and evaluated their analytics aspirations. Etleap prioritized Zava's most vital workflows, so that within the first few days priority sources and pipelines were up and running. These included webhooks to capture click data, which provide insight on how patients interact with Zava’s website, as well as indicate how well the company's products and marketing initiatives are working in their various operating locales. Additionally, Etleap encouraged Zava to run Etleap pipelines in parallel with existing Alooma pipelines and perform data parity checks. In their first days of Zava's Etleap migration, they were able to replicate all of their most important Alooma pipelines.

“Etleap’s direct support channel was a great asset throughout the entire trial and migration process,” says Bhattacharjee. “When I had an issue setting up a function in Python I talked directly with the engineer and got immediate help from him. We have had experiences with other providers where we were waiting 72 hours for a response, but this was much more responsive.”

Zava's data future bright after Etleap migration

Zava currently operates in three countries, has plans to expand into more markets, and recently partnered with Superdrug- one of the UK's leading health retailers- to better serve their millions of patients. Data is crucial to their innovative approach to healthcare, and when their ETL provider dropped support for the data warehouse they had come to depend upon, their future was uncertain. When they discovered that Etleap offered the best of Alooma's features and then some, plus unparalleled support throughout the migration journey, they secured their data future. They are now building novel and robust data pipelines that will help people find better health outcomes, and Etleap will be providing support at every step along the way.

“If there is any company looking for an ETL solution, have them give me a call and I'll tell them all about Etleap,”  concludes Bhattacharjee. “Great product. Great management. Great support.”

When Zava found out Alooma was dropping AWS support, they were blindsided. With three months to save their data pipelines, they needed an ETL provider that was AWS friendly. Etleap was built on and exclusively for AWS -- in fact, the only destinations Etleap supports are Redshift and Snowflake data warehouses, and S3/Glue data lakes. Etleap is built using AWS building blocks like EC2, S3, RDS, DMS, and EMR. For example, data processing is handled using EMR clusters that scale automatically in response to demand, and data is stored in S3 buckets for durability and high availability. This means users can reap the benefits of a robust AWS infrastructure without getting bogged down in low-level implementation projects.

What you can expect when migrating from Alooma to Etleap

Step I: Migration Assessment

Etleap will conduct a free migration consultation with your data team to gather requirements, including: sources needing support, existing Alooma pipeline configurations, and your analytics aspirations. During this call we determine your most challenging/vital pipelines (most complex transformations/most data) to map out a migration plan, and determine milestones we aim to hit during a POC.

Step II: POC and Migration

Etleap’s Alooma migration process is centered around recreating your existing Alooma pipelines on the Etleap platform. In the first days of the POC we get your highest priority sources and pipelines up and running in Etleap. We will enable you to run Etleap pipelines in parallel with existing Alooma pipelines to do parity checks. An Etleap engineer dedicated to the POC and migration will check in frequently throughout the process.