Blog
Five custom tools worth building, and why most are not.
Almost every list under this heading is a roundup of software to subscribe to, usually with affiliate links. This is the other kind: tools you build on your own site, which is a much smaller category and a much better investment when it works. It frequently does not work, so the failure mode is here too.
Why build one at all
A tool does two things a written page cannot. It answers a query whose intent is to do something rather than to read about something, and it gives people a reason to link to you that does not involve asking them to.
That second one is the real argument. On a site with no authority, the constraint is almost never that you have too few pages. It is that nothing links to you, and that is what a new domain is actually waiting on. A tool is one of the few assets that earns links passively, because people cite what they used.
The five patterns
Ordered by cost, cheapest first. Each of these has been built at least once rather than reasoned about.
| Pattern | What it is | What it earns | Cost |
|---|---|---|---|
| The generator | Takes a few fields and hands back something the visitor copies out: markup, a config file, a tag. | Ranks for "X generator", which is a query a written page cannot satisfy. The most reliably linkable of the five. | Low. Mostly a form and a template. |
| The checker | Measures the thing people argue about, and shows the measurement rather than a verdict. | Settles a recurring argument, which is what gets it cited. Weak on links unless the measurement is unusual. | Low to moderate, and rises fast if the measurement has to be accurate rather than approximate. |
| The calculator | Turns "it depends" into a number the visitor can act on, usually pricing or sizing. | Qualifies enquiries before they arrive and removes the reason to bounce to a competitor who published theirs. | Low to build, high to agree internally. The argument is never technical. |
| The finder | Turns a database you already have into an answer to a question with intent behind it. Near me, in my size, for my case. | The strongest commercially. Targets high-intent queries no static page can answer. | High. This is a real build with real data problems. |
| The comparator | Puts two or three things side by side with the relationships already worked out. | Removes the decision that was sending people away to compare elsewhere. | High, and the data modelling is harder than the interface. |
Generators are the best first build
The pattern is a form and a template, and the payoff is out of proportion to the effort. Somebody searching for a schema generator or an hreflang generator wants to produce a thing, and there is no article that satisfies that intent. You are not competing with content, you are competing with other tools, which is a much shorter list.
The ones on this site are the schema markup generator, hreflang generator and robots.txt generator. All of them run entirely in the browser with nothing uploaded, which is worth doing for its own sake and is also the thing that makes people comfortable recommending them.
A checker earns its place by taking a position
The meta description checker here measures pixel width rather than counting characters, on the argument that Google truncates by rendered width and a character limit is a convenient fiction. That stance is the reason the tool exists: a fortieth tool counting to 155 would have no reason to be built. If your checker measures the same thing as everyone else's, it is a page with a form on it.
The finder is where the money is
The highest-value build of the five, and the only one that regularly justifies itself commercially rather than through links. The pattern: you already hold structured data, and there is a high-intent query it can answer that no static page can.
A near-me tool built for a client site is the clearest example, targeting a query with substantial monthly volume that a written page could never satisfy. The interesting engineering was not the distance maths, it was making the location layer self-maintaining, so a static map falls back to cached term meta and then to a one-off lookup. New locations work with no code change, which is the difference between a tool and a liability. It was server-rendered rather than built client-side, for the obvious reason that a tool a crawler cannot read earns nothing.
The part these lists leave out
Most custom tools fail, and they fail the same way. They get built, they get published, nothing links to them, and they become an expensive page with a maintenance cost. The tool was fine. The distribution never existed.
Two specific traps. The first is orphaning: a tool sitting in the sitemap with no contextual internal links pointing at it has no internal authority and reads as peripheral. That is a real mistake made on this site, where two studies sat on two inbound links each for a week, and it is worth wiring the links at publish time rather than afterwards.
The second is building client-side only. If the output requires JavaScript execution and the page ships empty HTML, you have built something Google has to come back for on its own schedule. Render the useful state on the server and enhance it.
How to decide
Two questions. Is there a query where the intent is to do rather than to read, and do you already hold the data that answers it? Two yeses is a build worth costing. One yes is usually a page.
And be honest about the third question, which is who is going to tell anyone it exists. A tool with no distribution plan is the same mistake as publishing a page per keyword, just with a larger invoice attached.
Keep reading
More from the blog
Technical SEO in 2026, in priority order
Four questions in strict sequence, not a flat checklist of ninety items with no order to them.
Read itBest WordPress SEO plugins for 2026, compared
All of them output the same tags. The differences are real and they are not differences in SEO.
Read itHow to improve Core Web Vitals in 2026
The three metrics broken into their parts, the fixes that move them, and when to stop.
Read itKeep reading
Related reading
Book me
Got an idea for a tool?
Describe the query you want to answer and the data you already hold. You get an honest read on whether it is a build or a page, roughly what it would cost, and whether anything would realistically link to it.