Proofseal

Origin

Origin

Proofseal did not start from "I want to build a SaaS." It started as work for my thesis.

While studying for the Master of Science in Decision Making and Applied Analytics at Minerva University, I kept thinking about a very practical question:

When we are handed an analysis that looks complete, how much evidence is enough to carry it into a real decision?

The question did not appear suddenly. It accumulated slowly, from the classroom to work and back into research.

  1. Learning

    Evidence · Modeling · Formal Reasoning · Scientific Inquiry

    In coursework I began to deliberately separate "what we observed" from "what this evidence can actually support."

  2. Practice

    CloudAD · Marketing · Data · Measurement

    Budgets really move because of a number, so I found myself pausing before the answer to ask what conditions and assumptions it depends on.

  3. Complex Decisions

    plantārium · Operations · People · Context

    Many operational decisions have no ready-made dataset: data matters, but data always lives inside a context.

  4. Research

    MDA Thesis · Measurement Validation

    The thesis turned these doubts into a researchable question: what checks does a measurement result need before we know how far its evidence currently reaches?

  5. Proofseal

    Keep one more layer of review between a measurement result and a business decision.

Learning to look at answers differently

In the MDA program I met decision making from several angles. Evidence-based decision making taught me to separate what we observed from what the evidence can support. Statistical modeling made me look closely at data, variables, confidence intervals, and the assumptions behind every result. Formal methods asked another question: if a judgment matters, can we take the reasoning apart so that the conditions, assumptions and logic are all visible, instead of keeping only the final answer? Together these courses settled one thing for me: a good decision is not about gathering the most data or the prettiest number. It is about whether we know what we know, why we know it, and what we still do not know.

CloudAD: when measurement really moves budgets

At CloudAD the questions were very concrete. Clients would ask: which channel is working? Where should the budget move? How far can we trust what ROAS, attribution or a model estimate is telling us? Platforms produce many numbers and models produce many answers, but I found myself pausing before the answer more and more often: which data conditions does this result depend on? Is the method suited to the question at hand? Which assumptions would change the answer if they failed? Are we looking at correlation, attribution, or evidence that can actually support a causal claim? None of this is a rejection of measurement. Precisely because measurement moves real resources, it deserves to be examined more carefully.

plantārium: decisions are never just a model

In operations at plantārium I faced a completely different system: space, people, sustainability, education, events, revenue, costs, and stakeholders who all affect one another. Many decisions had no clean dataset to answer them. Sometimes the data was insufficient; sometimes goals conflicted; sometimes a decision that looked reasonable on paper stopped being the same answer once real people, time and context were added. That experience gave me another layer of understanding: data matters, but data always lives inside a context. What needs to be visible is not only the result, but the limitations, the assumptions, the uncertainty, and the decision the answer is about to be used for.

The thesis sharpened the question

For the thesis I tried to turn these standing doubts into a researchable question. What I wanted to study was not just "which measurement method is better" but the layer before it: how much should I trust this measurement result for an actual business decision? In other words: what checks does a measurement result need to go through before we know how far its evidence currently reaches? Over time, that question became Proofseal.

So Proofseal appeared

Proofseal did not begin as a commercial brand. It was the research and product prototype I started building for the thesis, bringing together questions that usually live in separate places: is the data sufficient to answer this question? Is the method suited to the context of use? Which assumptions does the result depend on? What evidence supports the conclusion, and what evidence limits it? What checks has the model or measurement been through? What kind of decision is this result about to support? In the end, none of this needs to be compressed into a single "trust score." Proofseal keeps verdicts at different levels: Supported, Conditional, More evidence needed. Sometimes the most useful result is not a yes or a no, but knowing: this is how far the evidence currently reaches.

From a thesis to a system people can use

Proofseal still carries its original research question. It is not out to prove any platform wrong, not built to replace any measurement method, and not meant to decide for its users how budgets should be spent. What it tries to do is simpler: keep one more layer of review between a measurement result and a business decision, so that the data, methods, assumptions, evidence and limitations can all be seen before the decision is made. That is where Proofseal starts.

Proofseal is an independent research and product project. Minerva University is part of its academic origin and does not endorse or certify the product.