Competition is only half the picture.

I've worked on competitive analysis for years, and for most of that time the job looked much the same. You mapped the landscape: who the rivals were, how they were positioned, what they charged and how they packaged it. You built feature matrices and sales battlecards, ran win/loss interviews and read the analyst reports. Then you refreshed it every quarter, or every year.

It worked because the landscape moved slowly. A serious new competitor took years to build, and you usually saw it coming.

Porter's Five Forces always said rivalry was only one of five pressures on a business. Two of the others are the threat of new entrants and the threat of substitutes. For a long time they moved slowly enough to leave in the background. They don't any more.

Building software has become cheap and fast. That changes where competition comes from.

• Personal softwareCustomers build their own with AI coding tools, and cancel. A weekend of work can replace a subscription.
• CopycatsAn open-source alternative with a following can pull users away. It doesn't have to be better. It only has to be free, good enough and talked about.
• AI agents & skillsThe job moves inside the assistant your customer already uses: an MCP server, a Claude skill, a ChatGPT plugin. They never open your product again.
• New entrantsA small team can launch a credible product in weeks, not years, and be funded before you've noticed.
• Platform absorptionWhen a big platform ships a feature that does your job, a product that was a thin layer on top can lose its reason to exist overnight.

None of these shows up on a competitor list until it's already costing you customers.

How strong is your moat? Warren Buffett popularised the idea of an economic moat: whatever stops competitors from taking your customers. Hamilton Helmer's 7 Powers names the durable kinds: scale economies, network economies, counter-positioning, switching costs, branding, a cornered resource and process power.

The question has always mattered. What's new is how quickly a moat can erode. Features can be copied in a weekend. What holds is the harder stuff: data others can't get, deep roots in how customers work, real switching costs, network effects, distribution you own, and the pace you ship at.

Moatly measures these from public evidence, and keeps them apart from the pressure on them. A strong moat under heavy pressure is a different situation from a weak one in a quiet market, and each needs a different response.

One open-source repo is noise. Three in a quarter, each with more stars than the last, is a trend. You only see trends by watching every week.

When you can see one coming, you can respond in time:

1. 01Double down on distribution and marketing.
2. 02Ship the features that keep you ahead.
3. 03Change your price or your commercial terms.
4. 04Deepen integrations and the reasons customers stay.
5. 05Build the community a clone can't copy.

A few changes a year used to be enough. In an AI market, it isn't.

Keep your competitive analysis. Your known rivals still matter, and the ways you track them still work. Moatly watches for the rest: the weekend projects, clones, AI assistants, new launches and platform moves that never make it onto the list.

It reads public sources (your own site and its history, GitHub, Hacker News, package registries, the MCP Registry, new domains, company filings and the big platforms' release notes), keeps only the genuine threats, and tells you how sure it is of every reading. Nobody can pay to change one.