Experiment

Website Asset Clusters

A live experiment testing whether a role-defined cluster of specialized websites can outperform a single all-in-one site on resilience, asset reuse, and monetization flexibility.

Krisada Eaton Updated September 11, 2026 Ongoing 211 views
Hypothesis

A website asset cluster built around specialized roles (hub, supply, collections, marketplaces, discovery, and exit) will create more durable search visibility and more monetization options than a single-site model built in the same niche.

Measurement Frame

What is being tracked.

Status Ongoing
Started April 25, 2026
Updated September 11, 2026
Duration Ongoing
Branches 6
Setup

How the test is structured.

The experiment is anchored in a fourteen-domain visual asset build spread across multiple stages of maturity. Some properties are live, some are mid-rebuild, and some still exist only as domains with strategy documents. That unfinished state is part of the test rather than a flaw in it.

The core variable is specialization. Instead of asking one site to handle storytelling, inventory creation, curation, commercial search, top-of-funnel discovery, and eventual asset resale, the cluster assigns those jobs to different properties.

The system is being observed across six operational dimensions: how many assets can be classified and routed cleanly, how often one piece of source inventory can be repurposed across multiple endpoints, whether discovery traffic can be separated from transaction traffic, whether marketplace pages convert more cleanly when upstream sites handle education and curation, whether the network creates stronger topical authority than isolated sites, and whether mature properties become independently sellable or leasable.

Observed Signals

What has happened so far.

September 11, 2026 checkpoint ... the architecture can scale. The performance claim is still open.

Three implementations now give the idea something real to push against.

The Green Campaign turned a contributed domain collection into thirteen live properties across six brands in four days. The IBM campaign applied the same role-specialized logic to eighteen properties inside one commercial subject. BellyUp ATL now runs six properties from one shared entity layer containing 31 verified restaurants across 19 neighborhoods.

That proves reuse, specialization, and repeatable deployment. It does not prove superior search performance or monetization. BellyUp's discovery impact has not been measured yet. The AS400System branch also exposed the risk concentration can hide: twelve related hosts lost Google visibility together when the domain family was reclassified.

The cluster was operationally resilient. It has not yet proved search resilience.

Decision Thread

What would support or challenge it.

Status: architecture supported, performance hypothesis open.

A website asset cluster is no longer just a diagram. The portfolio has reproduced the model across green-living, IBM infrastructure, and local restaurant discovery. One entity store can feed several useful surfaces, and one subject can support clearly distinct specialist properties.

The harder claim remains unproven: that the cluster outperforms one concentrated site on discovery, revenue, and standalone asset value. The next checkpoints need routed human use, conversions, independent citations, and sale or lease evidence, not another domain count.

Extended Analysis

The deeper read.

This experiment tests more than SEO. It tests operating model design.

Most website builders make one of two mistakes. They either over-concentrate value inside a single site, or they over-expand into multiple sites with no defined system logic. Website asset clusters are the middle option: expansion with assigned roles.

The measurement framework is intentionally broader than rankings. A cluster can create value through asset reuse, cleaner audience segmentation, better-fit monetization, stronger AI legibility, and transferable site equity. If those gains fail to materialize, the model should be treated as overhead, not innovation.

Experiment Lab

Keep Following the Tests

Move from this open thread back into the full experiment library.