The brief that almost had an obvious answer

The Selves of Us is a literary magazine dedicated to the interior lives of African women. Its editorial vision is precise: not a content platform, not a blog network, not a newsletter aggregator, but a magazine with the formal ambiguity of that word intact. A curated, edited, designed publication that happens to live on the web, with all the authority that implies and none of the compromises that most web publishing forces.

When the brief arrived, the obvious answer was Ghost. Ghost is a beautiful product. It is well-engineered, thoughtfully designed, and has evolved significantly over the past several years. For a publication that wants to publish and grow quickly, Ghost is a defensible choice. The same is true of Substack for audience-building, or a well-configured WordPress install for something more traditional. These tools work. Millions of publications use them successfully.

We built from scratch instead. Not because the off-the-shelf options were bad. Because none of them were The Selves of Us.

What platform dependency actually costs

The cost of a publishing platform is not only the monthly subscription. It is the sum of every editorial decision the platform forecloses. When you publish on Substack, your content lives on Substack domain structure, served from Substack infrastructure, styled within Substack design constraints, discoverable through Substack recommendation algorithm, and portable only to the extent that Substack decides it should be. The subscription fee is the visible cost. The invisible cost is everything your publication cannot be because the platform has already decided what it is.

For a magazine with a strong visual and editorial identity, these invisible costs compound quickly. The typeface is not the one you would choose. The spacing between lines is not the one that serves the prose. The image handling is not the one that would honour the photography. The navigation is the one every Substack publication has, because Substack is not interested in your navigation. It is interested in its network.

This is not a criticism of Substack or Ghost. It is a description of what platform dependency means. A platform built for many publications cannot be optimised for yours. It can only be optimised for the average of all publications it serves. If your publication is not average, you are paying a permanent tax on your identity to use it.

What we built instead

The Selves of Us platform is 93 pages of custom Next.js application. It has a custom content management system, a reading experience designed for long-form prose, an issue management structure that mirrors traditional magazine publishing, a newsletter engine, and an editorial workflow that the team can operate without a developer. Every decision was made in service of the publication identity rather than a platform constraints.

The typeface is Cormorant Garant, chosen specifically for its handling of long-form literary prose. The line length is controlled to approximately 65 to 75 characters, the range that reading research consistently identifies as optimal for comprehension. The image treatment is full-bleed where it serves the work and restrained where the prose needs space. The navigation is quiet and does not compete with the content for attention.

None of this required novel engineering. It required the freedom to make the right decision for this publication rather than accepting the decision a platform had already made for its average user.

24 articles in 90 days

The platform launched in the first quarter of 2026. In the first 90 days of operation, The Selves of Us published 24 articles. This number is significant not because of its size but because of what it represents: an editorial team using a platform they own, without workarounds, without fighting constraints, without needing developer support for routine publishing operations.

The editorial workflow is entirely self-service. New issues are created in the CMS. Articles are written, edited, and reviewed in the CMS. Images are uploaded and cropped in the CMS. Publication schedules are set in the CMS. The developer role after the platform was handed over was exactly zero: not because the team does not need support, but because the platform was built to be operated by the people who actually work on the publication.

This is the standard that custom publishing infrastructure should meet. Not the developer can do it in an afternoon. Not you can submit a support ticket and wait. The editorial team can do it themselves, today, without asking anyone.

The true cost comparison

A custom publishing platform costs more to build than a Ghost subscription. There is no avoiding this. The question is not whether the upfront cost is higher. It is whether the total cost over the lifetime of the publication is lower, and whether the value delivered by platform ownership exceeds the value that would have been delivered by a platform subscription.

For The Selves of Us, the calculation is clear. The platform is owned outright. There are no monthly fees that scale with subscriber count. There is no dependency on a company whose business model may evolve in ways that are not aligned with the publication interests. There is no limit on what the platform can become as the publication grows, because the platform is a codebase that can be extended rather than a product that can only be configured.

The upfront investment was real. The ongoing cost is infrastructure only. And the publication has an asset it will own for as long as it chooses to operate: a platform built exactly for it, that no one else has, that reflects its values in every detail from the typeface to the database schema.

This is what platform ownership actually means. Not a technical achievement. A publishing philosophy, implemented in code.