Jelle De Zwaef

The through-line

Across nearly every post, Jelle De Zwaef returns to the same gap: the difference between what an organization pays for and what it actually uses or understands. Whether it is FME licenses sitting idle [1][9], a handover document that "does not solve" the loss of an outgoing engineer [3], or a BIM attribute that passes visually but fails an OTL compliance check [12], his preoccupation is with the quiet drift between the paperwork and the workflow underneath it. This has sharpened rather than shifted over time: early in his career as an IT consultant he lost sleep over problems that were not his to solve [11], and running Nordend has turned that same attentiveness into a business model built around auditing the gap before it becomes a crisis [1][9][3].

On renewal audits

De Zwaef treats the FME renewal cycle as the moment truth gets told or avoided. He recounts auditing a Belgian agency with "23 FME desktop licenses" of which only "9" were active, renewed for four years "by a procurement process that nobody had questioned" [1]. His framing of the reseller's job is explicit: "A volume vendor sells you 23 licenses next year. A partner-reseller asks why you are paying for 14 unused ones" [1]. He extends this into a standing offer of discovery calls for "teams approaching an FME renewal in the next 90 days," aimed at producing "a clear view of which licenses are earning their cost" [9]. He also holds resellers accountable for representing clients back into the vendor relationship, noting that his team submitted "a detailed bug report to Safe Software" that later shipped in a release, and that "you are paying the partner margin for nothing" if your reseller never does the same [15].

On patience

De Zwaef names Hemingway, Humboldt, and De Gerlache as his heroes, not business leaders, because "the pattern across all three is patience" [4]. He explicitly pushes back on industry culture: "Our industry talks about pace as if speed were the only kind of progress" [4]. The work he is proudest of at Nordend "did not happen in a sprint" but over "a year of slow, low-volume conversations" [4]. His line on the cost of speed is blunt: "Speed has a cost. The receipt arrives months later in the form of workarounds nobody planned" [4].

On asking for help

He describes asking for help as "the hardest skill I have had to learn in years of running Nordend," using bookkeeping as the concrete example [5]. He still keeps "an hour each month on the books" not to do the work but so "the company can tell me how it is doing without using words" [5]. He generalizes this into a tension he sees in founders broadly: those who never ask for help "drown in the work," while those who "hand it all off lose touch with what is actually changing" [5]. This connects to his account of ownership as a founder: seven years into Nordend, "the stakes are higher" than his consulting days yet "I sleep better than I have in twenty years," because "ownership closes the loop" [11].

On institutional knowledge

De Zwaef treats a resignation letter as "the most fragile moment in any spatial data team," since the departing engineer "remembers why every workaround exists" and "a 40-page handover document does not solve this" [3]. His response is a structured pairing exercise: "One Nordend engineer pairs with your outgoing expert before they leave," documenting "the routing logic, and the reason a specific transformer was chosen over the obvious one" [3]. The same worry about undocumented drift appears in his description of BIM-to-GIS translation work, where unresolved OTL attribute failures accumulate into "four workarounds nobody documented" by the third project [12], and in municipal geoportal pipelines where "nobody remembers what changed" once a file has passed through several hands [13].

From the stage

In the Solutions Review interview at FMEUC, De Zwaef describes founding Nordend three years after already working with FME technology since 2006 across cartographic software, public safety, and utilities roles, and he details a specific service his company built to help customers deploy FME Server processes across development, test, and production environments with proper governance, a response to organizations struggling with process stability whenever environments changed [16]. This governance-across-environments problem is not addressed anywhere in his written posts, which focus instead on licensing, handover, and translation-layer issues.

Takeaways

• Before an FME renewal, audit which licenses are actually active and what the underlying workflows do, not just whether the license still works [1][9].
• Treat a resignation on a data team as an urgent two-week window to pair the outgoing expert with a successor and document routing logic and transformer choices, not just write a handover document [3].
• Expect BIM-to-GIS and OTL translation failures to compound silently across projects if workarounds go undocumented; catch them early with a focused discovery session on the specific failing attribute [12].
• Keep a small personal touchpoint on your own numbers or workflows, even after delegating, so you notice drift before it becomes obvious [5].
• If you buy FME through a reseller, ask what bugs or roadmap changes they have personally pushed to Safe Software on your behalf [15].
• Reduce platform noise deliberately; De Zwaef left Twitter and reports his industry knowledge got sharper once fewer hours went to hot takes and more to focused reading [2].