The Covenant

We built Observatory for people who already know what good work looks like — the career publisher or the business operator who built an audience they refuse to treat as a pipeline. You understand your audience not as a metric to move, but as people who chose you: readers, viewers, customers, members, a community formed around something specific they care about. You didn’t need a platform to tell you that. You needed one that understood it too.

Most analytics tools treat your work as a distribution problem. More traffic, better conversion, stronger engagement rate. These goals aren’t inherently broken. But they’re incomplete. When they become the sole frame of reference, the real question stops being asked: what does your audience need from you, and are you giving it to them? Observatory exists to hold that question open.

The system has a logic to it. Publish more. Reach more. Watch which signal goes up, give it more of that. The platforms reward this, the tools measure it, and the treadmill keeps moving. The problem is not that this produces bad work — sometimes it produces excellent work. The problem is that after a while, the prior question stops being asked: is this still where we want to go? Observatory is built for the moment you want to ask that question with data behind you, not just instinct.

Here is what we will do:

1. Surface the patterns in your channels that you cannot see from inside any one of them.
2. Show you where your audience is already moving — in search, in conversation, in the topics they return to — and help you understand what that movement means relative to what you’ve built.
3. Give you the same kind of analytical clarity that large organisations pay teams of people to produce, calibrated to the scale of your work.

We will do this in plain language, with reasoning attached, so you can argue with it.

Here is what we will not do: generate your content. Observatory will surface what the signals point toward — patterns in your data, gaps your audience is telling you about, topics with strong evidence behind them. The recommendation is ours; the work itself is yours, entirely. Not a draft, not a headline, not a paragraph meant to go live under your name. We use AI for analysis and synthesis only — to read patterns across your data and surface implications you might have missed. All of that analysis runs on your workspace’s own data, according to priorities you set, and you can turn it off. We are not a content machine. We are a reading instrument.

The embedded Analyst works from your workspace data. That is a principled choice, not a gap in the product. The value it gives you — precision, trustworthiness, reasoning you can follow back to a source — depends on a defined scope. An analyst that reaches for the internet to fill in what the data doesn’t say is not being more helpful; it is being less reliable. When you need your workspace data read alongside external research, the Observatory connects to your LLM of choice. That is where those two instruments meet — your data, brought by Observatory, and the open web, reached by your existing AI tools. The composition is yours to make.

There is a version of analytics work that AI handles competently: here is the signal, get more of those. Observatory is built for the work that comes before and after that — the question, the judgment, the decision to move toward something different. We use AI for analysis. The thinking is yours.

We will not optimise you toward goals that conflict with your mission. If the data suggests you could gain traffic by covering something that has no business being in your work, we will not flag it as an opportunity. The judgment is yours. Observatory will not blur that line.

We will not surface cross-site behavioural tracking data — no HubSpot tracking pixel, no LinkedIn Insight Tag, no Meta Pixel, no equivalent infrastructure that follows individual visitors without their active awareness. This is not a missing integration. Observatory understands your audience through what they choose to engage with, and that distinction is the foundation of the trust that makes the intelligence worth having. Surveillance data and signal data are different things. We will not treat them as the same.

Reviews and attributed public comments are received speech — given to be heard by the person who wrote them. Observatory surfaces them to return that voice to you, not to build profiles or rank individuals.

We will not treat direct visits as a gap. When your readers share your work through private channels — Slack threads, WhatsApp conversations, email forwards, closed community spaces — the visit arrives without a referrer, because the channel doesn’t pass one. Most analytics tools flag this as unmeasurable traffic and try to work around it. We read it differently: this is your audience trusting your work enough to bring it somewhere private, to people they actually talk to. That is not a measurement failure. That is evidence of the kind of reach that doesn’t show up in a ranked source list because it moves through relationships, not platforms. Observatory will show you how much of your traffic arrives this way. It will not try to resolve it back to individual sources, infer the platform it came from, or help you instrument around it. The signal is that it happened. That is enough.

We will not make you feel behind. The frame we work in is: here is what your audience is telling you, and here is what you might do about it. The emphasis is always on what you might do — not what you should have done three months ago. Foresight is meant to help you see what comes next, not audit what you’ve missed.

When you tell Observatory what direction you are moving in — what success looks like from the inside — that statement stays in your workspace. It is stored only to be read back against your data, by you, in the context of your own work. It is never aggregated with the statements of other workspaces. It never reaches any model training pipeline. It is never used for any purpose outside your workspace’s own reflection loop. Reads of your stored direction by Observatory’s analysis tasks are logged — you can see when your direction was read, and why. Deletion is complete and permanent, including from backup storage, and takes effect the moment you request it. Every surface in Observatory that quotes your stored direction shows you where it came from, when you wrote it, and a link back to the original. There is no invisible interpolation of what you said into what the product tells you. The interpretation is always attributable. The original is always visible.

We come to this work with years of experience building and running organisations where the audience relationship is the product — across media, content, community, and place — the kind of experience that teaches you what your audience needs before the data confirms it, and what it costs when you get that wrong. Observatory is built from that sensibility. When we say we understand the difference between an organisation with a point of view and one that’s just trying to rank, we mean it from the inside.

The people who find their way to Observatory usually weren’t looking for analytics. They were looking for a way to make decisions that matched how they think about their audience — and a way to know, over time, whether they’re still going in the direction they chose. If that describes you, you’re in the right place.

We believe that understanding your audience deeply, across all your channels, is an act of care. Observatory is built to support that work. Nothing else.

— No Media Company