llms.txt
Compare with other sites' llms.txt files →# Convex Works
> Convex Works is a data systems firm based in Amsterdam, serving Europe. We help growing companies build one complete data system they own and trust — accurate, compliant, and feeding every tool and decision.
Companies come to us when they can't trust their numbers, when useful first-party data leaks between their website, product, warehouse, ad platforms, and reporting, or when they want to own the data their business runs on instead of renting it inside vendor silos.
We work in three stages:
- **Audit** — We trace how customer and revenue data moves through your stack and where it gets dropped, delayed, or trapped, then return a prioritised list of fixes and a build plan.
- **Build** — We build collection, storage, transformation, and delivery on infrastructure you control. GDPR and CCPA are handled in the data model and pipelines rather than patched on later, and richer first-party signals feed the ad platforms, AI workflows, sales tools, and dashboards you already pay for.
- **Grow** — We stay on as an embedded data team, maintaining the system as vendors and schemas change and turning new business questions into shipped datasets, analyses, and automations.
Contact: hello@convex.works. Sitemaps: https://convex.works/sitemap.xml and https://convex.works/sitemap-figures.xml.
## Core pages
- [Home](https://convex.works/): What Convex Works does, who it's for, and how an engagement works.
- [About](https://convex.works/about): Why Convex Works exists, how the team thinks about owned data systems, and who founded the company.
- [Tools we work with](https://convex.works/tools-we-work-with): How we work with website platforms, analytics, customer data tools and APIs to connect acquisition, product activity and data the business owns.
- [Press kit](https://convex.works/press-kit): Download Convex Works logos, icons and profile pictures, individually or together in one ZIP.
- [Free data audit](https://convex.works/free-audit): Apply for a free audit. We review your public tracking in about a week and return a report on data gaps, leakage, and priority fixes.
- [Contact](https://convex.works/contact): Request an intro call or send an inquiry.
## Case studies
- [Minuttia](https://convex.works/case-studies/minuttia-research): How Minuttia studied 115,078 AI answers with Convex Works, followed by a second study on self-promotional content.
- [HouseFacts](https://convex.works/case-studies/housefacts): How first-party measurement recovered conversions hidden by Safari tracking prevention and ad blockers.
- [Pneuma](https://convex.works/case-studies/pneuma): EU-hosted, consent-first measurement of the onboarding funnel and marketing, with conversion counts kept honest.
- [Viral Loops](https://convex.works/case-studies/viral-loops): A first-party data system feeding one shared event layer, with budget-deciding conversions confirmed server-side.
## Blog
- [Blog](https://convex.works/blog): Writing from the team at Convex Works.
- [10 Marketing Analytics Agencies in October, 2026](https://convex.works/marketing-analytics-agencies): A comparison of ten agencies, their marketing analytics services, published client work and prices, with our selection criteria and disclosure.
- [We built tide to host our own video calls.](https://convex.works/why-we-built-tide): Nikitas Filosofof and Filippos Taprantzis on building and open-sourcing the self-hosted video call software used at Convex Works.
- [Why and how we created Convex Works](https://convex.works/why-and-how-we-created-convex-works): Nikitas Filosofof on starting Convex Works and choosing to focus on data systems.
## Research
- [Research index](https://convex.works/research): Studies by Convex Works and our partners.
- [How AI search cites self-promotional content.](https://convex.works/research/self-promotional-content): Our work with Minuttia on how AI answers cite vendors’ own product recommendations, using 18,084 software questions across six search and AI platforms, with a link to the full report on Minuttia.
- [AI Search Visibility Data Study](https://convex.works/research/ai-search-visibility): Our work as Minuttia's data partner on collection and analysis across 24,067 prompts and five AI search surfaces, with a link to the full report on Minuttia.
- [European software data study](https://convex.works/research/b2b-eu-data-systems): An analysis of 2,357 European B2B and B2B2C software companies covering websites, customer-data tools, consent banners, privacy policies and advertising.
- [Study methodology](https://convex.works/research/b2b-eu-data-systems/methodology): The study's selection criteria, evidence sources, analysis methods and limitations.
- [ArcUno: A Topology-Sensing Tangible Toolkit for Lean Knowledge Graph Modeling](https://convex.works/research/arcuno): Filippos Taprantzis’s Creative Technology bachelor’s thesis at the University of Twente and Vrije Universiteit Amsterdam, supervised by Victor de Boer. A tangible toolkit that turns physical nodes and edges on a desk into an RDF/Turtle knowledge graph, evaluated with six domain experts, with a link to the full thesis.
- [Contracting for initiative: How governance shapes decisions](https://convex.works/research/contracting-for-initiative): Nikitas Filosofof’s University of Amsterdam master’s thesis, inspired by client work at Convex Works. An online experiment on employee and contractor recommendations, with 231 participants in the main analysis and a link to the full thesis.
## Mentioned in
External references linked from the About page:
- [Viral Loops — How to Create a High-Quality Referral Program (5-Step Guide)](https://viral-loops.com/blog/how-to-create-high-quality-referral-program-guide/): Referral-program guide by Nikitas Filosofof.
- [University of Amsterdam — Contracting for initiative: How governance shapes decisions](https://scripties.uba.uva.nl/search?id=c15619613): Nikitas Filosofof’s master’s thesis, which discusses Convex Works as the inspiration for the research (English).
- [Universidad Politécnica de Cartagena — Los neobancos y su estrategia de comunicación online](https://repositorio.upct.es/server/api/core/bitstreams/7c6357bf-e7d2-4f95-aecd-da8750077b51/content): Undergraduate thesis by Julián Rabal Conesa on neobanks and online communication (PDF, Spanish).
- [University of Piraeus — Η επόμενη ημέρα των παραδοσιακών τραπεζικών ιδρυμάτων - Ευκαιρίες / απειλές και η επίδραση της διοίκησης ολικής ποιότητας](https://dione.lib.unipi.gr/xmlui/bitstream/handle/unipi/16557/Spyros_ΜΔΕΟΠ1929.pdf): Master's thesis citing Nikitas Filosofof's Viral Loops case study of Revolut (PDF, Greek).
- [Ghost — Talking the talk](https://ghost.org/resources/talking-the-talk/): Newsletter quoting Nikitas Filosofof on affiliate marketing.
- [GrowthWaves — AI search visibility study for SaaS](https://www.growthwaves.com/p/ai-search-visibility-study): Article by George Chasiotis on AI search visibility for SaaS, based on 24,067 prompts.
## Careers
- [Careers](https://convex.works/careers): There are no open positions at the moment.
## Team
- [Nikitas Filosofof](https://www.linkedin.com/in/filosofof/): Co-founder of Convex Works.
- [Filippos Taprantzis](https://www.linkedin.com/in/ftaprantzis/): Co-founder of Convex Works. Publications on [Google Scholar](https://scholar.google.com/citations?user=EE6E8zgAAAAJ).
## Optional
- [Privacy Policy](https://convex.works/privacy-policy): How Convex Works handles personal data.
<!-- data-study-figures:start -->
## European software data study — figures
Each figure below is published under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Cite the figure page.
Every figure page carries the chart as a PNG, the underlying numbers as an HTML
table, and ready-made plain-text and BibTeX citations.
### Marketing
- [Advertising channel combinations across the cohort](https://convex.works/research/b2b-eu-data-systems/figures/ad-channel-combinations): Each company is counted once, under the channels where we found its ads.
- [Advertising-result and website-tag overlap by channel](https://convex.works/research/b2b-eu-data-systems/figures/website-tag-alignment): Website tags and ads often appeared separately on all three platforms.
### Privacy disclosure
- [Share of tracking sites without a consent banner](https://convex.works/research/b2b-eu-data-systems/figures/bannerless-tracking-share): Gauge showing that 44.3% of sites using tracking technologies did not show a consent banner.
- [Consent journeys and tracker activation](https://convex.works/research/b2b-eu-data-systems/figures/consent-journeys): The chart shows whether each website contacted a known tracking domain before the visitor responded. It also shows whether the website contacted a new one after acceptance.
- [Data system score compared with estimated traffic](https://convex.works/research/b2b-eu-data-systems/figures/data-system-score-traffic): Points are horizontally jittered; DS scores themselves are discrete. The 1,116 sites produce 83 unique scores: 37.2% score exactly 40, and 80.6% occupy the ten most common scores. Traffic is shown as log10(1 + estimated traffic), preserving all 211 zero-traffic sites. Spearman ρ = -0.026: higher-traffic sites do not appear to have better or worse DS scores. The published chart data contains only true DS score and rounded log-transformed traffic—no company names, domains, labels or hover data.
- [Privacy policy disclosure gap by tool](https://convex.works/research/b2b-eu-data-systems/figures/disclosure-gap): Not one of the ten tools appeared in most privacy policies. Even Google Analytics, which was named most often, was missing more often than it was mentioned.
### Competition
- [Employee-count distributions for European and American companies](https://convex.works/research/b2b-eu-data-systems/figures/company-size-swarm): This comparison is based on reported employee counts for study companies and accepted direct competitors in Europe and the United States. Each dot is one company; companies without a reported count are not shown.
- [European and United States competitor shares by product vertical](https://convex.works/research/b2b-eu-data-systems/figures/competition-balance): Categories are ordered from the largest US proportion to the largest European proportion.
### The cohort
- [European countries represented in the study](https://convex.works/research/b2b-eu-data-systems/figures/country-distribution): This map is based on the country recorded for each of the 2,357 companies in the final study group.
- [Company founding year compared with estimated organic traffic](https://convex.works/research/b2b-eu-data-systems/figures/founding-year-traffic): This comparison is based on 1,897 companies for which Apollo reported a founding year, compared with Ahrefs’ estimate of monthly organic traffic.
### Data trade deficit
- [Whether customer data can stay in the EU](https://convex.works/research/b2b-eu-data-systems/figures/processing-and-location): Each cell represents one of the 361 tools reviewed.
- [Supplier terms for secondary use of customer data](https://convex.works/research/b2b-eu-data-systems/figures/supplier-terms): Tools whose terms gave no clear answer are not included in the bars.
- [Customer-data tool detections across company websites](https://convex.works/research/b2b-eu-data-systems/figures/trade-balance-observations): A tool is counted once for each website where it was found, so one website can contribute several observations.
### Customer-data tools
- [15 most common combinations of visible customer-data tools](https://convex.works/research/b2b-eu-data-systems/figures/tool-combinations-upset): The top bars show how often each combination was found. The left bars show how many websites had each tool. Connected dots show which tools make up each combination.
- [Distribution of visible customer-data tools per company](https://convex.works/research/b2b-eu-data-systems/figures/tool-count-distribution): The final bar groups websites with ten or more tools.
<!-- data-study-figures:end -->