Cervello and the expensive art of getting data to agree

The revealing number in one Cervello project is 22. That was how many minutes a particular analytics query took to run. Along the way, it spilled 1.86 terabytes of data - a technical inconvenience with a financial consequence. After optimization, the query took 12 minutes and produced no spillage. Somewhere inside the magnificent promise of the cloud, somebody still had to inspect the plumbing.

This is a useful way into Cervello because the data business is exceptionally good at describing itself in clouds. Transformation. Intelligence. Insights. The nouns float beautifully. The actual work often concerns a query, a spreadsheet, or two departments using the same word to mean different things. Cervello earns its place in the conversation by dealing with those particulars.

The bill arrived before the revelation

The query belonged to a revenue cycle management organization using Snowflake for ordering and scheduling analytics. Users wanted fresh information. The annual budget was 50,000 credits; forecast consumption exceeded 100,000. Cervello reports resizing and splitting warehouses, rewriting inefficient queries, selectively clustering tables, and reducing capacity during quieter periods. Overall credit usage fell by a reported 20 to 35 percent.

Credits are units of platform consumption, not a universal dollar price. Nor does the published reduction alone reconcile every part of the original budget gap. The useful conclusion is narrower: architecture and query behavior can materially change operating costs. Buying capacity is only one response to a slow system. Understanding what the system is doing can be another.

A consultancy with a screwdriver

Founded in 2009, Cervello developed a business spanning enterprise performance management, data management, business intelligence, and customer relationship management. These are neighboring territories with a common border dispute: whose data should determine what the business does next? The firm’s work includes cloud integration, custom applications, and ongoing managed services.

In January 2019, A.T. Kearney acquired the company. The attraction included systems that would continue working after a consulting assignment ended. There is a commercial distinction here. A recommendation can be accepted in a meeting. A functioning application has to survive Monday morning.

“Winning with data, that is what it’s all about.”Alex Liu, Kearney’s acquisition announcement, 2019

Cervello’s market position sits between executive advice and technical delivery. Its parent supplies industry and strategic expertise; Cervello contributes architecture, engineering, analytics, and development. Large consultancies and specialist implementation firms also compete in that territory. The interesting distinction is the specific combination within Kearney, rather than a claim that nobody else can connect a database.

The business model follows the work: enterprises hire a team to advise, implement, develop, or operate a solution. Snowflake, Anaplan, and Salesforce are among the platforms used. A buyer should distinguish the software subscription, the implementation assignment, and the future operating bill. They pay for different things, even when the sales conversation makes them sound like a single purchase.

The parent had its own spreadsheet problem

Kearney’s 2023 account of its internal planning transformation supplies an unusually candid customer example. Its process involved emailing approximately 50 spreadsheets worldwide. Finance and HR used separate headcount tools, creating discrepancies. Cervello helped implement Anaplan with common data, automated feeds, and shared definitions. After the initial headcount and profit-and-loss work, Kearney expanded into long-range planning.

The revealing change was an agreement about what the numbers meant. A shared system made that agreement consequential. This is the practical appeal of connected planning: a staffing choice and a financial forecast can be discussed together. The software makes a connection possible; people still have to decide how their assumptions belong together.

A reading of the approach across Cervello’s cases, not a proprietary product workflow.

A shopping problem disguised as a data problem

At an unnamed tools manufacturer with more than $15 billion in revenue, departments acquired external data separately. Purchases overlapped, and access was fragmented. Cervello built a marketplace with Collibra at the front and Snowflake behind it, initially covering seven domains and twelve sources. It was made accessible to more than 60,000 employees. That describes eligibility to use it, rather than a count of active shoppers.

The delightful detail is the shopping experience. Companies already understand catalogues, discovery, and reusable inventory when the subject is physical goods. Applying that logic to datasets makes an abstract asset easier to find. The underlying management problem is equally familiar: without visibility into what the organization owns, different people can pay for the same thing.

This suggests a question readers can copy before launching a large data program: what do we already have, and who can find it? The answer may require engineering. It may also require a conversation with procurement.

The arithmetic needs its small print

For an insurance, finance, and holiday services provider, Cervello assessed a fragmented data landscape and designed a consolidated architecture and implementation plan. Its published business case describes a £1 million investment that could deliver £6 million in value over three years. The word “could” matters. It is projected program value, not a public statement of cash already saved or a standard Cervello fee.

For a chemicals company, the work extended to a prioritized portfolio of more than 150 data initiatives, governance, and a learning curriculum for over 130 employees. That mix is telling. An organization can have plenty of ideas and still need help choosing their order. Architecture gives projects somewhere to live; prioritization decides which deserve to arrive first.

These assignments fit businesses with multiple systems, competing functional needs, and enough complexity to justify outside expertise. A small team with one reliable database may have a much simpler problem. Equally, a company unwilling to assign owners or change workflows can commission an elegant platform and leave the old habits intact. Those are practical implications of the cases, rather than guarantees about any future engagement.

Where the old name leads now

By October 2026, Cervello’s website directs visitors to Kearney Activate, and its LinkedIn page says it is retired. The current destination presents a wider collection of data, AI, engineering, and transformation capabilities. Those broader practice results should not automatically be booked as Cervello achievements.

The older materials remain useful for understanding the expertise. Salesforce privacy services, for example, are described in three scopes: Advisory, Essentials, and Premier, ranging from advice to hands-on configuration and more complex implementations. The pattern is familiar: make a business requirement concrete, configure the tools, and train the people who will inherit them.

A prospective customer can borrow that discipline. Specify the decision that needs improving, the data behind it, the person responsible, and the expense that should change. Then ask what will be working when the consultants leave. Cervello’s best public stories are persuasive because their answers are so ordinary: a faster query, a discoverable dataset, a forecast that two departments can finally discuss without first arguing about the file.