Our client was a Fortune 500 US independent broker-dealer with over 17,500 financial advisors and over 1tn USD in  advisory and brokerage assets. They had a large application estate with nearly 1,000 applications they had either developed, bought or acquired through mergers and takeovers. The applications were captured in ServiceNow CMDB but there was little knowledge around flows, owners, data, and batch jobs.

Additionally, the client also wanted to roll out a new data strategy. Part of this engagement with their business community was to educate and inform about the data strategy and its impact on their work.

We were asked to implement an open source enterprise architecture tool called Waltz. Waltz had been originally developed at Deutsche Bank and had recently been released as open source software under FINOS (Fintech Open Source Foundation). Waltz is not widely-known in financial services yet and we saw this as a great opportunity to demonstrate the benefits of using open source tools.

To support the data strategy rollout, the client asked if we could build a simple and clear internal website to show the new data strategy and data model. The data model would be navigable to drill-down into more detail and provide links to existing documentation.

Our approach:

With our extensive implementation experience, we put together a small, experienced, cross-border team to deploy and configure Waltz. We knew that understanding the client’s data was key; what data was required, where was it, how good was its quality. Waltz uses data around:

  • Organisational units – different structures depending on the viewpoint (business, technical)
  • People – managerial hierarchies, roles, responsibilities
  • Applications – owners, technologies, costs, licences, flows, batch jobs
  • Data – hierarchies, entities, attributes, definitions, quality, owners, lineage
  • Capabilities – owners, services, processes
  • Change – initiatives, costs, impact

We split our work into a number of workstreams:

  1. Data readiness – understand what data they had, the sources, and the quality
  2. Data configuration – understand the relationships between the data and prepare it for Waltz
  3. Waltz implementation – understand the base open source version of Waltz with its limitations, gather the client requirements (like single-sign on and configurable data loaders), develop the features into Waltz, and deploy Waltz at the client
  4. Data strategy website – understand the audience, design website prototype options for client review, build an interactive React website for the rollout roadshows

The project was challenging because, as ever, the state of the data. There were multiple inconsistencies which hinders the use of tooling to bring order. We needed to identify those inconsistencies, see who should own them, and ensure they were resolved.

With the flexibility of an enterprise architecture tool, it was important to be clear around the specific problems we wanted to solve for the client. We identified 10+ potential use cases that we worked with the client to narrow down. Future extensions of the project enabled us to extend into these other use cases.

One such problem was around batch job documentation. The client had thousands of Word docs specifying batch jobs transferring data between internal and external applications. These documents were held in SharePoint, Confluence, and local drives. This made it difficult to find information about specific batch jobs if something went wrong, for example.

We used the applications captured in Waltz and linked them together. We developed a new data loader that could import Word docs and extract the batch job information automatically from them. This was used to populate Waltz and make this information searchable, reducing the time spent by Support teams to find out about failed jobs.

One common negative that is raised about similar applications is the effort involved to get data into the application. Waltz accelerates this by sending surveys out to crowd-source knowledge from across the organisation. We found this a great way of engaging with users and capturing their experience into Waltz.

Our results:

We were able to deploy an open source enterprise architecture tool on a client’s AWS cloud within three months. This included adding new features, such as single sign-on, improving existing Waltz capabilities, like the data loaders, and defining the data standards to enable smooth data integrations with source systems.

Using Waltz showed the client the value of bringing together disparate knowledge from around the organisation into one place. It does expose data gaps, but we always see this as a benefit for the client, as any improvement in data quality yields improved business results.