Notes on Precise and Serendipitous Creative Tools
The rough distinction between precision and serendipity can help develop better long term expectations for new creative tools.
Serendipitous tools tend to be beginner friendly and are valued for their happy accidents, something good that wasn't exactly what the user was aiming for. The quality of outcomes is less dependent on skill.
Precise tools can be difficult to learn but once the user gains confidence they might feel like they could use it to do anything. For the experienced user, outcomes are closely aligned with very specific intentions1, and they learn something whenever there's a surprise.
In the long term, precise tools tend to have better retention, whereas serendipitous tools usually have faster growth. These differences mostly come down to switching costs and network effects. For precise tools, the user incurs a high cost (in time and energy spent on skill learning) before they can start to accrue the benefits. That slows viral growth, but once a user is comfortable with the tool, newer alternatives start to look less appealing. More importantly, the tool tends to be low-level and open-ended, enabling users to invent new techniques, make YouTube tutorials, want to agree on an industry standard, etc. These things can lead to precision tools having strong network effects.
Serendipitous tools don't have the same learning curves. They're most useful at launch, but the more people use them, the worse they are at producing relatively novel output. So newer alternatives are usually more serendipitous than the older, more popular tools. The result is that both the network effects and switching costs can quickly become net negative.
Of course, these are not hard rules. A precise tool can have faster viral growth, sometimes simply through hype. A serendipitous tool can have better retention with continuous innovation.
The distinction isn't strictly binary either, any given tool can be a mix. But most tools are much better at one than the other. Why would that be the case? It seems common for people to use both in parallel on any given project, wouldn't it be convenient to have them packaged together?
Not necessarily. Users typically want to replace their precise and serendipitous tools at different frequencies, so it makes sense for these use cases to be solved by separate products.
Knowing how specific your intentions are is much less trivial than people often think. Indeed much of the initial struggle with precise tools can come from beginners not realising the ambiguity of their ideas.