The 200-line Chrome extension that replaced my spreadsheet

I used to export search term reports to CSV every Friday.

Then I'd open a spreadsheet. I'd copy my proprietary negative keyword library into another sheet. Then I'd manually cross-reference. Rows of garbage traffic against my internal blacklist. Marking false positives. Building a list of new negatives to add.

It took about three hours a week.

One day I got frustrated and asked myself: why am I doing this manually?

The answer: because I was the only person who needed it. Building a tool for a three-hour-a-week problem that only one person had seemed wasteful. So I just lived with it.

Then I realized: I could build a Chrome extension in about 200 lines of JavaScript that did this automatically.

The extension sits in my Google Ads UI. When I'm looking at a search term report, it injects a button. I click it. The extension grabs all the search term data from the table, runs it against my negative keyword library (which lives in a Google Sheet), and highlights false positives.

Green highlight means "this is a legitimate negative, add it." Yellow highlight means "this query is suspicious, probably negative, but worth a second look." Red highlight means "this looks like false positive, probably bad to add."

It saved me three hours a week. And it took about four hours total to build (including debugging and deployment).

Why I didn't build this sooner

The barrier was mental, not technical.

I knew how to build it. The logic is trivial. Grab table data. Loop through it. Compare against a list. Add CSS classes to highlight matches. Done.

But I never built it because the infrastructure felt hard. I'd have to:

• Set up a development environment
• Learn Chrome extension APIs beyond what I already knew
• Figure out how to deploy it
• Deal with versioning

All of that is boilerplate. None of it is the actual feature. So I just... didn't.

I lived with a three-hour weekly task instead of spending a day building a tool to eliminate it.

What changed

The same thing that let me build Langton Tools. Claude Code.

I described what I wanted: "Build a Chrome extension that runs on Google Ads search term reports. When a user clicks a button, it grabs all the search term data from the table, compares it to a list of negative keywords in a Google Sheet, and highlights matches."

Claude Code built it. First pass had a few bugs (the DOM selector needed tweaking), but it was 90% there. I told Claude what was broken. Five minutes later, it was fixed.

Total time investment: about 90 minutes.

Most of that was testing and making sure the DOM selectors were stable. The actual "building the feature" part was maybe 20 minutes.

Why this matters

This extension exists in the gap between "problem big enough to build a tool for" and "problem small enough to just live with."

Most work has dozens of these gaps. Tasks that repeat weekly. Tasks that take 2-3 hours. Tasks that would take a day to automate.

Usually you just live with them. Because the barrier to building the automation is high and the time savings are "only" a few hours a week.

But if you do this work for 10 years, those few hours add up to months of your life. Months spent doing something a computer could do for you.

The Claude Code tool flattens the barrier. Now the decision is simpler: "Is it worth 90 minutes to save 3 hours a week?" Yes. Obviously. Do it.

Before, the decision was "Is it worth 2-3 days to save 3 hours a week?" No. Obviously. Live with it.

What this unlocked

After I built this extension, I looked around and found about 15 more tasks that fit the same pattern. 2-3 hour repetitive tasks. Could be automated. Not quite big enough to feel worth building.

Now I'm building them all. Because the time investment is reasonable.

A script that automatically flags high-CPL keywords for review: 40 minutes.

A dashboard that shows budget pacing status: 90 minutes.

A tool that monitors cross-brand auction overlap: 2 hours.

All of these are things I would have written down as "nice to have" and never built. Instead, they're shipping in the next couple weeks.

And that's actually the origin story of Langton Tools. I started building one tiny extension to solve my own problem. Then I realized I could build 20 more. And that other practitioners probably had the same problems.

So why not package them and ship them properly?

The pattern

Here's what I learned from building this extension and a dozen more:

If you do a task more than three times, code a solution.

It doesn't have to be perfect. It doesn't have to be beautiful. It just has to save you time. And thanks to Claude Code, building it is way faster than the barrier used to be.

I have 15 years of spreadsheets in my history. Tasks I repeated hundreds of times because the automation felt too hard to build.

Now I look at anything that repeats more than quarterly and I build a tool for it.

The time I save is real. And it compounds. A three-hour-a-week task becomes an hour a month. Over a decade, that's months of your life back.

Not bad for a 200-line extension.