The institution thesis

What we are building, and why that question keeps mattering

Every organisation eventually has to answer the question of what it actually is. Not the pitch deck version. Not the category the market assigns it. The actual answer, the one that determines who you hire, what you build, how you allocate time and capital, and what you refuse to do even when the money is there.

For Mustard Seed Group, the answer is: we are building an institution.

That word gets used carelessly, so let me be precise about what it means here and what it does not mean.

An institution is not a large company. Size is not the point. An institution is an organisation structured around a long-term mission rather than around a single product cycle, a funding round, or an exit. It accumulates standards, culture, and capability over time in a way that makes it more useful and more coherent the longer it runs, not less. Universities, certain publishers, certain studios, certain families of companies: these are institutions. Most startups are not, and they were never intended to be.

MSG is not a startup. It is not optimising for an exit at year five or year seven. It is not structured to be acquired, merged, or dissolved when conditions become inconvenient. The institution thesis is the deliberate choice to build for a 20-year horizon rather than a five-year one, and to let that choice drive every consequential decision from the beginning.

The distinction between a holding company and a venture fund

People sometimes reach for the term "holding company" to describe what MSG is. It is a reasonable shorthand. The group holds equity in, and exercises operational authority over, a set of distinct operating businesses. On paper, that looks like a holding structure.

But holding companies and venture funds share a logic that does not apply here. In a fund, the portfolio exists to generate returns distributed across a large number of bets, on the assumption that most bets will underperform and a small number will carry the whole structure. The fund has a fixed life. It has LPs who expect capital returned. The individual company matters less than the portfolio math.

MSG is not that. There are no external LPs. There is no fixed life. The operating companies in the portfolio are not bets in a diversification strategy: they are projects within a single coherent mission. The mission is increasing human capability, and each business in the portfolio addresses one dimension of that mission. Orbit addresses the capability to build and operate products. TUXX addresses the capability to deploy custom intelligence for specific commercial contexts. Benediction Lab addresses the underlying research questions that the operating businesses will eventually need answers to. All Purpose and CheekyGains address the dimensions of performance, culture and community that shape what people believe they can do.

These companies share strategy, standards, infrastructure and identity. That is not a holding company structure. It is closer to what you see in the great design studios or the great publishing houses: a group of differentiated outputs united by a single sensibility and a single set of standards, all answering to a common mission rather than to a common return target.

What changes when you design for 20 years

The most important thing a long time horizon changes is the hiring logic.

In a five-year frame, you hire to fill a capacity gap. You need someone to write code, manage a product, run campaigns. You define the role around the immediate need, hire the person best suited to fill it quickly, and move on. That is rational for a company that may not exist in seven years and needs to ship something in six months.

In a 20-year frame, every hire is a cultural decision first. The person who joins MSG in 2025 may well be a significant figure in what MSG becomes in 2035. The values they carry, the standards they hold themselves to, the way they treat hard problems and uncomfortable truths, these compound over time. A single poor cultural hire at an early stage can take years to correct, and the correction is expensive not just in time but in the erosion of standards that is very hard to recover once it has started.

This is why we move slowly on people, even when the immediate need feels urgent. Urgency is usually a short-term frame masquerading as reality. The long frame says: hire deliberately, hold a high standard, and accept short-term capacity constraints rather than compromise on the quality of the people inside the institution.

The second thing a long horizon changes is the product philosophy.

Short-term product thinking rewards features that drive immediate engagement and short-term retention. The metrics that matter in a five-year frame are active users, monthly revenue, gross margin, and the numbers you will put in front of an acquirer or a Series B investor. These are real things. But they are not the whole picture.

Long-term product thinking asks a different question: is this product making the person who uses it genuinely more capable? Not more dependent on the product, not more engaged in the platform's own terms, but actually more capable in the world. The distinction is not academic. A product that creates dependency rather than capability can show excellent short-term metrics while producing long-term harm. A product that genuinely increases capability may have a slower adoption curve but produces the kind of deep loyalty that sustains a business for a decade rather than a funding cycle.

This is the logic behind Naira inside CheekyGains. An AI performance coach could, if designed with short-term engagement metrics as the primary driver, optimise for making the user feel like they need Naira's input constantly. That would look good on a dashboard. It would be wrong. The right design is one where Naira makes the user a better athlete, a more self-aware performer, someone who over time needs less coaching rather than more, because the quality of the coaching has genuinely transferred to the person. The revenue model has to follow from that logic, not fight against it.

The third change is the relationship to capital.

An institution does not structure itself around the next raise. Capital is a resource, not a scoreboard. The discipline is to understand what each unit of capital can build and to make sure the build is worth doing, not to maximise the amount raised as a proxy for progress. There are moments where external capital accelerates something important, and those are worth pursuing. But the decisions about what to build and how to build it have to precede and drive the capital strategy, not be determined by what capital is currently available to fund.

Culture and standards as the durable assets

The things that are hardest to build are the things that last longest and matter most.

Culture is not a set of values on a wall. It is the accumulated behaviour of an organisation under pressure. What do people do when a decision is genuinely difficult? What standard do they hold the work to when no one is looking? How do they treat one another in the parts of the working week that no one documents? These behaviours, repeated over years, become the actual culture. And culture, once established, is extremely hard to change. That cuts both ways: it is very hard to recover a good culture after it has been allowed to deteriorate, and it is very hard to shift a bad culture once it has calcified.

The implication is that standards have to be non-negotiable from the beginning. Not performatively high, but genuinely and consistently high. This applies to the quality of written work, the quality of code, the quality of product decisions, the quality of how we treat clients and collaborators, and the quality of how we handle situations where the honest answer is less comfortable than a convenient one. You cannot build an institution on a foundation of convenient answers.

Standards also have to be specific. Vague aspirations to "excellence" produce nothing. Specific standards, enforced consistently over time, produce the kind of institutional quality that becomes self-reinforcing. When the standard is clear and the people who hold it are trusted and respected, new members of the organisation learn the standard not through a manual but through proximity to people who live it. That is how culture actually transfers.

What MSG's institutional identity looks like in 2025

We are early. There is no pretending otherwise. The companies in the portfolio are at different stages of maturity. The research agenda at Benediction Lab is still finding its edges. The consumer businesses are still proving their theses in markets that reward clarity of purpose. The commercial studio work at TUXX is still establishing its reputation for a particular kind of rigour.

What is already true, in January 2025, is the frame. We know what we are building and we know the time horizon we are building on. We know what the portfolio is for and what holds it together. We know the standards we intend to hold and the kind of culture we are building towards.

The institution thesis is not a destination. It is a design principle: build things that compound, make decisions at the time horizon that matches what you are actually building, hold the people inside the organisation to a standard that will still look right in ten years, and keep the mission clear enough that every individual project can be evaluated against it rather than against the short-term noise of a given moment.

This is the work. The point is not to talk about it. The point is to do it well enough, for long enough, that the institution speaks for itself.