Baxnet Blog · founder-note
The Gap Between Collecting Your Life and Understanding It
Short answer: Personal data is not automatically useful because it has been collected. The missing layer is flexible analysis that lets people ask their own questions of their own life.
A person can have the whole archive and still not have the answer.
That is the strange part about personal data. We talk a lot about access, export, dashboards, charts, and ownership. Those things matter. But they do not automatically help someone understand what happened in their own life.
You can download years of messages and still not know when a relationship started to feel different. You can collect health readings and still not know what pattern is worth paying attention to. You can save every receipt, route, note, photo, and timestamp, then end up with something that feels less like self-knowledge and more like a storage problem.
The data is there. The useful question is still hiding.
That is why a 2021 paper by Jimmy Moore, Pascal Goffin, Jason Wiese, and Miriah Meyer is such a useful reference point for the kind of product world Baxnet cares about. In “Exploring the Personal Informatics Analysis Gap”, the researchers argue that personal informatics has paid a lot of attention to collecting data and producing insights, but less attention to the flexible analysis people need in between.
Their phrase for this is the “personal informatics analysis gap.”
It is a small phrase with a large product implication.
The Missing Middle
Many personal data tools are good at capture. They know how to count steps, log sleep, store messages, record spending, preserve location, or create a timeline. Some are also good at producing fixed summaries: weekly averages, top contacts, trend lines, category breakdowns.
But real personal questions rarely arrive in fixed shapes.
Someone does not always ask, “What was my average?” They ask things like:
- When did this start changing?
- Is this a one-off or a pattern?
- What was happening around that time?
- Which examples should I look at first?
- Is the report missing something obvious?
- Can I trust this conclusion?
Those are analysis questions. They need room for exploration, correction, context, and judgement.
The paper studied how people with asthma engaged with personal air quality data over time. The details are specific to that domain, but the broader finding travels well: people had varied goals, explored data in different ways, sometimes discovered things through play or accident, and were not always eager to use analysis tools on their own.
That last point matters. It suggests the answer is not simply “give users more powerful tools.” Power without guidance can become another burden.
Personal Insight Needs A Workbench
For Mimoto, the same idea shows up around message history.
An export is not enough. A beautiful chart is not enough either. If someone is trying to understand a relationship, a difficult group chat, or the shape of a conversation over time, they need a kind of workbench for their own history.
Not a black box that declares what their life means.
More like a private place to ask better questions: show me the turning points, let me inspect the examples, let me change the period, let me compare this person with that group, let me see the source material, let me disagree with the interpretation.
That is the difference between a dashboard and a personal insight engine.
A dashboard shows what the system already decided to measure. A personal insight engine should help the person investigate what they actually care about, while keeping the raw material close enough that the answer can be checked.
This is also where privacy and usefulness meet. If the data is intimate, the analysis layer cannot be careless. It should make the person more capable without making them feel watched by the product. It should help them notice patterns without pretending to own the truth about them.
The Real Product Challenge
The deeper lesson from the analysis gap is that personal data products should not stop at possession.
Access matters. Portability matters. Local storage matters. But the next question is more human: can the person do anything meaningful with what they now have?
If not, we have only moved the confusion from someone else’s server to the user’s hard drive.
For Baxnet, this is one of the reasons personal insight engines feel like a distinct category. The goal is not to collect more of a person’s life. The goal is to help them make careful use of the parts they already have, under boundaries they can understand.
That is quieter than the usual AI promise. It is also harder.
Because the useful product is not the one that says, “Here is your data.” It is the one that helps the person ask, “What can I now understand?”
Further reading: Moore, Goffin, Wiese, and Meyer, “Exploring the Personal Informatics Analysis Gap: ‘There’s a Lot of Bacon’” (IEEE TVCG / VIS, 2021). The study centers on asthmatics and personal air quality data, so the point here is a design analogy for personal insight systems rather than proof about message-analysis users.