Why a hook library is worth keeping
Most teams that run ads have learned a lot about what makes people stop scrolling. The trouble is where that knowledge lives. Some of it sits in an ad dashboard that only shows the last few months. Some of it sits in a slide from a quarterly review. Most of it lives in the head of whoever ran the campaigns, and it leaves when they do. So when it's time to write the next batch of ads, the team starts from a blank page and guesses again.
A hook library fixes that. The hook is the first line of an ad, the part that has to earn the next second of attention. The angle is the reason to click that the hook opens, such as a problem the reader has, an outcome they want, or an objection they need answered. A good library records which hooks and angles beat your usual results, and for whom. It turns every ad you've ever paid for into a head start on the next one.
Three things separate a useful library from a swipe file of ads someone liked:
- Every entry has evidence. It shows the result the hook earned, the measure it was judged by, and the dates it ran. "This one did great" is an opinion. "This hook brought trial sign-ups at a lower cost than any other line in that ad set over six weeks" is something you can build on.
- Every entry has an audience. Restaurant owners and dental practice managers might both buy your bookkeeping software, but they respond to different openings. The library should tell you which hooks won with which people.
- The losers stay in. A retired angle with a note on why it lost saves you from paying to learn the same lesson twice.
This work suits an AI agent well. The record has to be updated after every round of ads, the results have to be read carefully so that a small sample doesn't crown a false winner, and the library has to be checked every time someone writes a new ad. Those are exactly the steps busy people skip.
Kite is an AI marketer. It researches your business, builds a growth strategy, and ships the work that brings you new customers. Here's how it keeps your hook library.
How Kite keeps a library of your best ad hooks
Kite keeps the library in your team's knowledge base, the shared record it builds when you sign up and adds to as it works. The paid advertising section holds what your ads have taught you: which angles won, which were retired, which are still being tested, and the period each round covered.
The library grows in a loop. Kite reads how your ads performed through your connected ad accounts. It compares each line against your usual results, using the measure you care about, and sets aside anything that hasn't run long enough to judge. It records the winners under the audience they won with and notes the angles it's retiring. Then, the next time you need ads, it reads the library first and writes the new batch from what has already worked.
- Decide what counts as beating your usual results
- Sort hooks by the audience they won with
- Fill the library from the ads you've already run
- Write new ads from the library
- Keep competitors' hooks separate from your own
Decide what counts as beating your usual results
A hook earns a place in the library when it does better than your normal results on the measure that matters to your business. That measure is usually cost per lead, cost per trial, or cost per purchase. Clicks alone can reward a hook that attracts curious people who never buy, so Kite judges by the measure you name. If you don't name one, it tells you which one it used.
Kite also refuses to call a winner too early. A line that only reached a few hundred people might look brilliant by chance. When Kite reads results, it marks lines that haven't had enough delivery to compare as untested and tells you the cut-off it used. Those lines stay in the library as "still under test" rather than being promoted or thrown out.
Here's an example. Say you sell project management software and run Meta ads to agency owners. Over six weeks, one hook, "Stop chasing timesheets every Friday afternoon," brought trial sign-ups well below the cost of your other lines in the same ad set. A second hook about client reporting earned plenty of clicks but almost no trials. Kite records the first as a winner for agency owners, with the cost, the measure, and the dates. It retires the second with a note that it drew clicks but not trials, so nobody mistakes it for a success later.
Sort hooks by the audience they won with
Kite files each winning hook under the audience it won with, because the same line can succeed with one group and fail with another. It works out the audience from how your campaigns are built and from the way each platform reports results.
On Meta, Kite reads the targeting on each ad set and can split results by age group, placement, and device. That's how it might find that a hook worked in Instagram Reels for younger buyers but not in the Facebook Feed. On LinkedIn, campaigns are targeted by job title, seniority, industry, and company size, so Kite can see whether a hook won with heads of finance or with operations managers. On Google, the audience is often defined by the search itself, so Kite notes the keyword theme or ad group where each headline earned its results.
Continuing the example, suppose you also run LinkedIn ads to operations leaders at construction firms. The timesheet hook does poorly there, but a hook about tracking crews across job sites beats your usual results. Kite records both outcomes. The library now tells you that the timesheet angle belongs to agency owners and the job-site angle belongs to construction, which is exactly the kind of lesson that gets lost when it lives only in two separate dashboards.
Fill the library from the ads you've already run
You don't have to start the library from today. Kite reads past performance from your connected Google Ads and Meta Ads accounts for whatever period you choose, so it can go back through months of ads and pull out the hooks that already beat your results.
If some of your history lives elsewhere, such as an old ad account, an export from a previous agency, or a spreadsheet where someone logged test results, upload it in the conversation. Kite records what it can verify from those files, notes the source of each entry, and marks anything without results as untested.
You can also tell Kite what you know. "Our discount angles never worked for enterprise buyers" is useful even without a spreadsheet behind it. Kite records it with your name and the date, so later readers can tell where it came from.
Write new ads from the library
The library pays off when you write the next ad. Before Kite writes a new batch, it reads the paid advertising section of your knowledge base. It finds the winning hooks for the audience you're targeting, writes fresh variations that keep the idea and change the wording, and leaves the retired angles out.
Each new batch also tests one or two new angles, so the library keeps learning instead of repeating the same three ideas forever. Every line in the batch is labeled with its angle and the source it came from, such as a past winning ad or a customer review. That labeling is what lets Kite read the next round's results back by angle and update the library again.
Kite writes the copy for Google search ads and for Meta, LinkedIn, TikTok, and X, and checks every line against the platform's length limits. It can also make the images and short vertical videos to go with the copy, in your brand's colors and fonts.
Your team launches the ads. On Google and Meta, you upload the batch in the ad platform. On LinkedIn, Kite can build the ads into a draft campaign in your account, and a team owner or admin launches it from Kite.
Keep competitors' hooks separate from your own
Plenty of teams keep a swipe file of other companies' ads, and it's a useful source of ideas. Kite's built-in ad research shows the ads your competitors are running on Meta, Google, and LinkedIn, including the copy, images, landing pages, and how long each ad has been running. An ad that a competitor has kept running for months is a reasonable sign that it's working for them.
Kite keeps that research apart from your hook library, because a competitor's hook hasn't been proven with your audience or your offer. It uses competitor ads to see which angles are crowded in your market and to find one nobody is using yet. Their claims stay theirs. Your library holds only what your own ads have earned.
Related reading
- Ad swipe file: real ad hooks worth saving, and how to build your own Coming soon Real examples of ad hooks worth saving and a guide to building a swipe file of your own.
What you get
A hook library in your knowledge base
The hooks and angles that beat your usual results, each with the result, the measure, the dates, and the audience it won with.
A record of retired angles
The angles that lost, with a note on why, so nobody pays to test them again.
A list of angles under test
The lines that haven't had enough delivery to judge yet, kept until there's enough data to decide.
New batches built from the library
Fresh variations of your winning hooks for the audience you're targeting, plus one or two new angles to test, with every line checked against the platform's limits.
Answers on request
Ask Kite which hooks have worked best with a particular audience, or why an angle was retired, and get an answer with the dates and results behind it.
How to get started
Tell Kite what you want the library to cover. You might say, "Go through the last six months of our Meta ads and tell me which hooks worked for which audiences," or "Build a hook library from our Google and LinkedIn ads and use it the next time we write ads."
Kite asks which measure you want hooks judged by, if it doesn't already know, and which audiences matter most. Then it reads your results, builds the first version of the library, and shows you what it found: the winners for each audience, the angles it would retire, and the lines that need more time.
You can work with Kite in your browser or in Slack. To read your own ad results, Kite needs your ad accounts connected. Without them, it can still research your competitors' ads and write from your customer reviews, but the library itself depends on results only your accounts hold.
Kite proposes the work, and your team launches every ad. Updating the library after each round doesn't need your approval, so it stays current without anyone remembering to do it.
Integrations used for keeping a hook library
Your ad accounts supply the evidence the library is built from. Kite's built-in tools handle the writing, the images and videos, and the research into competitors' ads, so the only connections you need are the accounts where your ads run.
Read which hooks earned results
Kite reads results for each campaign, ad group, and ad, finds the headlines and descriptions that earned conversions, and notes the keyword themes where they won. Your team adds the new ads in Google Ads.
Kite reads the text, images, and results of each Facebook and Instagram ad, along with each ad set's targeting. It can split results by age group, placement, and device to see which audience a hook won with.
Common questions
What is the difference between a hook and an angle?
The angle is the reason someone should click, such as saving time, avoiding a risk, or getting a result a customer reported. The hook is the specific opening line that delivers that angle. One angle can have many hooks, and testing several hooks for the same angle tells you how to say it, while testing different angles tells you what to say. Kite records both, so the library shows which reasons to click work for each audience and which openings deliver them best.
How does a hook library fit with the 3-2-2 method for Facebook ads?
The 3-2-2 method is a common way to structure a Meta creative test: three creatives, two primary texts, and two headlines, which Meta combines and learns from. The method tells you how to set up a test, and the library tells you what to put in it. Kite can fill the primary texts and headlines with variations of your proven hooks for that audience, plus a new angle or two, and record the results when the test ends. Kite writes Meta ads as packages of several primary texts with different openings and two headlines doing different jobs, so every combination reads well.
How long should a hook stay in the library?
A hook stays useful until the results say otherwise. Ads wear out as the same people see them again and again, so a hook that won last spring might be tired now. Kite keeps the date on every entry and reads fresh results each round. When a former winner stops beating your usual results, Kite notes the change instead of quietly deleting it, so you can see that the angle worked for a while and then faded.
Can I see the library myself?
Yes. Ask Kite in the conversation, in plain language, such as "Which hooks have worked best with construction firms?" or "Why did we retire the discount angle?" Kite answers from the library with the dates and results behind each entry. If you want something you can share with an agency or a new hire, ask for a report with a shareable link, which you can limit to people at your company.
Does Kite need customer reviews as well as ad results?
It helps. Ad results tell Kite which hooks worked. Customer reviews, testimonials, and sales-call quotes give it the words to write new ones. Kite builds every new line from evidence such as a past winning ad or something a customer said, and it labels the source beside the line. Your best hooks and your customers' own words can also feed your brand's voice guide, so the tone that sells in your ads carries into the rest of your marketing.
What other ad work can Kite do?
Kite writes and tests new ad variations, makes ad images and videos, and researches what your competitors are running. To set up a creative test from scratch, see Write ad variations for creative testing. To keep a wider record of what works across all your marketing, not just ads, see Keep a record of what works in your marketing Coming soon .