Twenty-seven years of handling other organisations' data

Mandate

Create the group's first data office, set one standard for how client data is separated, accessed, retained and deleted across every service line, and govern Lumivara's use of AI and its data products.

Outcomes

• Moved 138 client analytics environments into separate tenancies within fourteen months; 72 clients now hold their own encryption keys.
• Introduced monthly access statements for every sovereign-tier client after moving night operations to EU teams in Tallinn and Vilnius.
• Launched a contributor-only operations benchmark with 212 paying contributors and €8.4m revenue in 2025.
• Built an opt-in AI programme for service-desk tickets: 47 clients, 86% correct routing and a €6.1m saving in 2025.
• Established a register of 61 AI systems, each with a named owner, intended use and evidence of performance.

Decisions in context

The chief executive created the role after the deletion work in Sweden and Finland showed that each of our service lines made its own promises about client data and few could prove them. I began with an inventory of every place client data lived and every route by which our people could reach it. That inventory produced the most expensive decision of my career so far: separating our managed analytics service into one tenancy per client. Since then I have moved overnight administration of our sovereign tier inside the EU, written the terms on which we will reuse client data for our own models and set the rules for our data products. I also run our internal data platform, which supplies utilisation, delivery and finance data to the executive team. Our data protection officer reports to the chief executive rather than to me, and I report to the audit and risk committee each quarter on data protection, AI and client data risk.

Reason for the move

Current appointment; completing preparation for the EU AI Act's staged obligations and the Data Act's switching requirements.

Mandate

Integrate Brantmark's data platform business into Lumivara after the May 2019 acquisition and harmonise how the combined business held and deleted client data.

Outcomes

• Designed a single exit standard with per-client backup keys; the first deletion certificate was issued in June 2020.
• Kept 94% of the acquired business's recurring revenue through the first eighteen months after completion.
• Merged two subprocessor registers into one list of 41 suppliers, published to clients.
• Reduced backup retention for departing clients from thirteen months to 35 days.

Decisions in context

I came into Lumivara with the business I had helped build at Brantmark, and my first task was to make two hosting companies behave as one towards their clients. The commercial integration was largely planned before completion. The data integration was not. Our two companies used different backup schedules, different subprocessors and different exit procedures, and neither could show a departing client that its data had gone. A request from a Swedish public agency for a deletion certificate covering backups made the problem impossible to postpone. I led the work on per-client encryption keys for backups, a shorter retention period for departing clients and a signed certificate with an annual audit sample. The work delayed part of the integration programme by four months and cost €2.7 million. It also gave the chief executive a clear example of what a group-wide data standard could look like, and he asked me to build one.

Reason for the move

Appointed Lumivara's first group chief data officer when the role was created.

Mandate

Build a managed data platform business and prepare the company, as a processor for most of its clients, for the GDPR.

Outcomes

• Renegotiated about 1,100 client data processing agreements before May 2018.
• Built the company's first register of subprocessors, covering 34 suppliers.
• Grew data platform revenue from SEK 95m to SEK 180m in two years.
• Prepared the data room and data-handling evidence for the sale to Lumivara.

Decisions in context

After eight years of advising, I wanted to be accountable for the systems themselves. Brantmark was a founder-led hosting company with private equity backing, a good reputation among Swedish municipalities and almost no documentation of how it handled client data. I arrived as the GDPR's application date approached. Most of our clients were controllers and we were their processor, so the work was less about our own customers' personal data than about proving that we followed our clients' instructions. I led the renegotiation of about 1,100 processing agreements, built a subprocessor register and set the first rules for engineers' access to client systems. I also built a managed data platform service that nearly doubled in two years. When our owners decided to sell, I prepared the evidence that buyers asked for about our data handling. Lumivara's diligence team found the gaps we had already listed, which I think helped the sale more than it hurt.

Reason for the move

Brantmark was acquired by Lumivara Oyj in May 2019; I moved into a group role in the combined business.

Mandate

Deliver data warehouse and data platform programmes for Nordic clients, and from 2013 build the firm's practice in data governance and data protection.

Outcomes

• Led eleven data warehouse and platform programmes for banks, retailers and government agencies.
• Built a data governance and protection practice that grew from 6 to 45 consultants.
• Assessed transfers of personal data for 23 clients after the Safe Harbour arrangement was invalidated in October 2015.
• Completed a part-time master's degree in information technology law in 2013.

Decisions in context

I moved to Stockholm to work on larger data programmes than a regional operator could offer, and I stayed eight years. The early work was technical: data warehouses for banks and retailers, and a register data platform for a Swedish government agency. I became a partner in 2013, by which time clients were asking us less about building data platforms and more about governing them. I built a practice in data governance and data protection that grew to 45 consultants, and studied information technology law part-time so that I could read the contracts I was helping to write. The invalidation of Safe Harbour in October 2015 was the defining event of my last year. I spent the winter helping 23 clients find which of their systems sent personal data to the United States, and learnt how rarely organisations know where their data actually goes.

Reason for the move

Wanted to be accountable for running systems rather than advising on them; recruited by Brantmark Cloud.

Mandate

Build the operator's first data warehouse and use it to reduce mobile customer churn.

Outcomes

• Built a data warehouse combining billing, network and customer service data, live in 2005.
• Reduced annual mobile churn from 21% to 16% through retention offers based on service problems rather than calling patterns.
• Declined in writing to build a segmentation based on customers' most-called numbers; the chief executive upheld the decision.
• Introduced the company's first rules on analysts' access to traffic data.

Decisions in context

I returned to Vaasa to build something from the beginning. Rannikko had billing, network and customer service systems that had never been joined, and its mobile customers were leaving at more than a fifth a year. I built the data warehouse with a small team and used it to find which customers were leaving because of dropped calls, billing errors or slow repairs, rather than because of price. Retention offers aimed at those problems reduced churn by five points. Telecoms data also taught me where analysis should stop. Traffic records reveal a great deal about a person's life, and the confidentiality of communications is protected more strictly than most commercial teams expect. When marketing proposed segmenting customers by the numbers they called most often, I wrote a short note explaining why I would not build it. The chief executive agreed, and I have written such notes ever since.

Reason for the move

Recruited by Vessling & Aro in Stockholm for larger data programmes.

Mandate

Prepare, link and analyse register data for labour-market and education research under the institute's confidentiality rules.

Outcomes

• Produced the statistical analysis for nine published studies of employment, education and income.
• Helped rewrite the institute's disclosure rules for published tables after a withdrawn table in 2001.
• Built linkage and weighting procedures used by the register unit until 2010.

Decisions in context

Kallbacka was my apprenticeship in handling other people's data with care. The institute linked employment, education and income registers to study how people moved through the labour market, under confidentiality conditions that its researchers took seriously. I prepared and analysed the data for nine studies and gradually took on the linkage and weighting work for the whole unit. In 2001 I published a table in which one cell, for a small municipality, contained two people, and one of them was recognised. The table was withdrawn. My head of unit taught me statistical disclosure control that afternoon, and I then helped rewrite the institute's rules on minimum cell sizes and dominance for every published table. Those rules, adapted, now govern Lumivara's data products. I left after four years because I wanted to see whether the same discipline could inform commercial decisions, and a regional operator in my home town offered the chance.

Reason for the move

Returned to Vaasa to apply statistics to commercial decisions at a regional telecoms operator.