writing-top-performing-thought-leader-ads-tlas.md

Writing top-performing Thought Leader Ads (TLAs)

A TLA is served under a real person's name and looks like an organic post. The reader's instinct is "this is a person's genuine opinion" - not "this is an ad." The moment it reads like an ad, you've lost them. The goal is not to sell; it's to be so useful, honest, or interesting that the reader clicks through.

Apply the Anti-slop reference on top of everything here. Benchmarks below are from ZenABM's 2026 LinkedIn ABM Benchmarks Report (161,256 ads, $5.5M spend) - https://zenabm.com/linkedin-abm-benchmarks-report

The report's TLA rules (canonical summary)

✅ 1st person "I" voice - 65% of top performers (vs 30% bottom)
✅ Link at bottom - 75% place link in last 25% of text
✅ 1,000-1,500 characters - optimal length for value + attention
✅ Pain-point hooks - "I've seen so many..." pattern wins
❌ Avoid: webinar CTAs (0% in top 20), corporate "we" voice, hashtag spam

Required inputs - interview before writing

Never write a TLA without these. Ask the user explicitly:

Who is posting - name and position in the company. The voice, credibility and stories must be theirs.
For the specific use case this TLA covers: real numbers, personal experience, and learnings the author can share. A metric that's slightly uncomfortable to share, a mistake they made, a client story with figures, a before/after. If the user has none, push once ("even a rough number or one real anecdote"), then write a specificity-light draft and flag it as weaker.
The CTA destination - the client's demo or trial URL (a specific landing page beats a homepage).

Do not fabricate numbers, stories, or experiences. A TLA with invented specifics is a liability under a real person's name.

Non-negotiable rules

CTA: demo or trial link only. Never link to blog posts or content downloads. No "book a consultation", no "limited spots", no giveaways, no "comment X to get Y" mechanics.
CTA placement flexible, but always present - and put the link in the bottom 25% of the text (75% of top performers do; never mid-text - 40% of bottom performers).
Never mention or quote competitors. Describe the problem, not who else fails to solve it.
Write in 1st person "I" - 65% of top TLAs vs 30% of bottom. Corporate "we" voice: 0% of top performers.
Sound human. No three-bullet lists where every bullet starts with the same verb, no symmetrical sentence pairs, no corporate filler. Read it aloud before finalizing.

Benchmarks that shape the format

TLA median CTR 2.68% / weighted CPC-to-LP $3.06 - 77% cheaper per click than image ads ($13.23). Lead with TLAs.
1,000-1,500 characters is the sweet spot (55% of top performers). Under 500 = no value delivered; over 2,000 = loses the reader.
1-2 sentence paragraphs, line breaks everywhere (100% of top performers), 1-3 functional emojis (👍👇👉, ✅ for benefit lists), 0-2 hashtags.
Webinar CTAs scored 0% in the top 20 and 0.07-0.08% LP CTR. Never use them.
Specific free-trial offers ("37-day free trial") and FREE ungated resources appear across top performers.
Social proof with specific metrics appears in 45% of top TLAs: "$650,000 in pipeline in 90 days", "$12 in pipeline per $ spent".

The hook (first 2-3 lines - 90% of readers see nothing else)

Hooks that outperform:

Challenge an assumption: "It's tempting to believe there's a single source of truth for [X]."
Bold/contrarian statement: "[Popular tactic] is an oxymoron."
Results-first: "[UPDATE: specific number, outcome, timeframe]"
Personal narrative with a credible detail: "I made this mistake [N] times." / "I've seen so many businesses still running operations on Excel spreadsheets. It works, until it doesn't." (3.42% LP CTR)
Shared pain: "Every [job title] I talk to says the same thing."

Avoid: generic openers ("Running an agency is a constant balancing act" - 0.04% LP CTR), announcing the ad, starting with the company or product name, starting with a question, future-vision hooks ("As we stand on the cusp of 2026...").

Structure

The arc almost every high performer follows:

1. TENSION     - something is broken / misunderstood / underestimated
2. SPECIFICITY - a concrete example, data point, or personal story that proves it
3. INSIGHT     - the reframe: here's what's actually going on
4. RESOLUTION  - the better way (framework, tool, mindset shift)
5. PROOF       - result, number, or testimonial
6. SOFT CTA    - pull, not push, with the link

Tension and specificity are non-negotiable. Character budget: Hook 50-100 → Backstory 300-500 → Problem detail 200-300 → Solution + 3-4 ✅ benefits 300-400 → Social proof 100-200 → CTA + link 50-100.

Content formats that work: What/Why/How; Honest Results (include the number that's uncomfortable to share, admit what didn't work); "I Made This Mistake"; Contrarian Take (pick a side, ground it in something specific, attack the tactic not people); Named Framework.

Micro-moves from top performers: the pivot sentence ("But here's the part I keep coming back to."), the I've-been-there admission ("No shade. I've been there."), before/after year contrast, the punchy kicker that ties back to the hook.

The CTA

Pull, not push:

✅ "Want to find out more? Head to [link]" / "You can try it free for [X] days." / "Happy to share more if you're exploring similar challenges. [link]"
❌ "Click here to book a demo." / "Get started today." / "Sign up now."

Direct-command-at-bottom ("Read the full case study 👇 [link]") appears in 50% of top performers - direct is fine, salesy is not.

What consistently underperforms

Opening with the product name; feature lists without a problem frame; vague pain ("teams struggle with X"); long unbroken paragraphs; multiple CTAs; corporate tone / passive voice; purely positive posts with no admission of failure; bracketed links ("[Check out the report here - link]"); 2,000+ character tangents; 5+ hashtags; survey-data openers.

Reference examples

Live examples of the target quality (LinkedIn ad-preview links - open in a logged-in browser):

TLA brief format

Ad name: / ABM Campaign: / Ad set: / Persona:
Posted by: [Name, Title]
CTA link: [demo/trial URL]
Author inputs used: [the numbers/stories the user provided]

--- POST TEXT ---
[the full TLA, formatted exactly as it should appear]

--- WHY THIS WORKS --- (2-3 sentences: hook type, arc, benchmark patterns used)

Checklist before delivering

First line creates tension or makes a bold, specific claim
Written in 1st person, in the named author's plausible voice
At least one specific number, data point, or concrete example - provided by the user, not invented
No sentence over 25 words; paragraphs 1-3 sentences
1,000-1,500 characters (engagement-variant: 600-1,000)
Product appears as resolution, not centrepiece; no competitor mentions
Exactly one pull CTA, demo/trial link, in the bottom 25%
Reads aloud like a person, not a press release
Every sentence survives "could I delete this?"