Ask any serious museum curator what distinguishes a great private collection from an expensive one, and the answer is almost always the same: coherence. The works speak to each other. There is an argument, a sensibility, a sustained point of view running through every acquisition.
That does not mean narrow. Some of the most extraordinary private collections span centuries, media, and cultures — but they do so because the collector had a genuine and evolving intellectual position about what connects them. It is not about a theme in the decorative sense; it is about a way of seeing.
The practical implication for building a collection: buy what you are genuinely curious about, not what you think you should be curious about. Collecting driven by social signalling — acquiring names because they are the names to acquire — produces interchangeable collections that mirror each other across apartment walls and storage facilities. Collecting driven by genuine engagement with ideas produces something rarer.
This also has a market dimension. Collectors with strong points of view develop deep knowledge in their areas of focus. That knowledge is a competitive advantage. They recognise works that others overlook, understand pricing anomalies that reflect ignorance rather than value, and build relationships with artists and dealers who respect their seriousness. Over time, the collection and the collector's reputation reinforce each other.
Practical starting points: identify three to five artists whose work genuinely engages you, and spend six months learning everything available about them before buying anything. Read their interviews, follow their exhibition history, visit their shows. By the time you make your first acquisition, you will know more about the context of what you are buying than most buyers at a far higher price point.
The Meridian's AI is built for exactly this kind of sustained research — use it not just at the point of purchase, but as a tool for developing the knowledge base that makes a real collection possible.