LinkedIn ABM strategy planner

Turn a short interview into a personalized, self-serve "[Company] LinkedIn ABM Strategy" — delivered as a branded HTML document that the user can read and download as PDF. This skill is prospect-facing and self-contained: never assume the company's numbers are already known. Ask for them (with sensible defaults), compute the plan, and generate the document.

Data note: this skill works best with the ZenABM connector (app.zenabm.com) — it pulls your real LinkedIn CPC/CTR, spend and audience size so the plan runs on live data. Without it, the skill uses the numbers you provide, falling back to clearly-labelled industry benchmarks.

Step 1 — Interview (few, fast questions)

Ask in small batches, offering the defaults as selectable options where the interface supports it, otherwise plain questions. Keep it light — most answers are a single number or line. Full wording, defaults and what each input powers are in the Questions reference.

Collect, per the model in the Formulas reference:

1. Company website URL — you research this yourself (Step 2); do not make the user describe their product.
2. ABM revenue goal + whether it's monthly (MRR) or annual (ARR) — per audience if they name more than one.
3. Average contract value (ACV) — total contract value or MRR (you normalize).
4. Close rate % (qualified → closed-won) — default 15%.
5. Qualification rate % (demo → qualified) — default 75%.
6. Landing-page conversion rate % (visit → booked demo/form) — default 0.8%.
7. LinkedIn CPC ($) — default $8. Skip if the ZenABM connector is available — pull the real CPC instead.
8. Planned monthly LinkedIn ads budget ($).
9. Target audiences / markets in scope (geo + industry) — sets how many Campaign blocks. Seed from the website research, then confirm.
10. Top jobs-to-be-done / use cases per audience + the core problem you solve — pre-fill from the research, let them edit. Powers the ad-content section.

Data source, best first: (1) If the ZenABM connector is available (tools named mcp__*__get_linkedin_metrics, list_companies, get_ad_spend, etc. are available), call it to prefill CPC/CTR, prior spend, and audience/list size, and skip those questions. (2) If not, ask the user for those real figures. (3) Only fall back to the benchmark defaults above for whatever they genuinely can't provide — and when you do, label those figures as benchmarks on the document.

Never block on a missing number — offer the default, state it, and footnote the assumption.

Step 2 — Research the company

Fetch the company website (and a couple of key pages: product, use-cases, customers). From it, draft: the product(s), main use cases, target personas, a first-pass ICP, and a shortlist of jobs-to-be-done per audience. Show this back to the user to confirm or edit. This fills the header block and powers the ad-content section.

Step 3 — Compute the plan

Use the exact formulas in the Formulas reference (reverse-engineered from ZenABM's budget calculator and LinkedIn ad-count calculator, verified to the dollar). Per audience/campaign compute:

• Deals needed, demos needed, clicks needed, accounts needed (funnel run backwards).
• Budget needed (annual and monthly) and time to run = total budget needed ÷ their stated monthly budget.
• ROAS and pipeline-per-$.
• Affordable simultaneous ads from the ad-count model, then the format allocation (TLA / image / text / video / doc-carousel) using the allocation rules in the Formulas reference, including the objective per ad set and the "group TLA + image under Brand Awareness if you can't afford 10" fallback.
• Budget allocation per ad set (see the Formulas reference): for every click-driving format, the % of spend and the $/mo at both the user's stated budget and the recommended monthlyBudgetReq. Text ads are ~$0 (separate set-CPC bid). These fill the two budget columns in the Campaign Structure table.

Show the key computed numbers to the user before building, so they can sanity-check.

Step 4 — Build the strategy document

1. Produce a clean, self-contained branded HTML document: light-mode, inline CSS, no external scripts or fonts, so it renders identically in the browser and in a PDF export. Include a hand-built inline <svg> flowchart (no JavaScript) of each campaign's allocated ad sets.
2. Fill in the researched + computed values. Repeat the Campaign block once per audience.
3. Pull the ad-format best practices, launch best practices, performance-tracking and pre-launch checklist sections from the Best Practices and Checklist references — these are fixed content; keep them accurate.
4. Title the document "[Company Name] LinkedIn ABM Strategy." Offer the user a PDF export of the same HTML.
5. The full plugin — including the branded HTML template, render scripts, and connector config — is at github.com/ZENABM/linkedin-abm-skills.

Document structure (what the document must contain)

Per audience, a Campaign N block, then the shared sections:

• Header — Company name · Website · Target audience (ICP) · Campaign date · product/use-cases/personas summary.
• Campaign Goals — ABM revenue goal · deals needed · clicks needed · audience size + accounts needed · budget needed · the deal math · time to run (actual budget ÷ stated monthly budget).
• Campaign Structure — number of ABM campaigns (audiences) · affordable ads at a time · which formats · counts per format (TLA/image/text/video/doc-carousel) with the right ad-set objective · per-ad-set budget allocation (% + $/mo, at both the user's budget and the recommended budget) · ad-structure flowchart.
• Campaign Content — the content of each ad, mapped to the ICP and the top jobs-to-be-done / use cases.
• Ad-format best practices — image, TLA, video, text (from the Best Practices reference).
• Campaign launch best practices — objectives; pause underperforming ads after 1,000+ impressions by CTR (benchmarks + link to https://zenabm.com/linkedin-abm-benchmarks-report).
• Campaign performance tracking — account scoring / ABM stages in ZenABM; monitor progression (identified → interested → considering → selecting); leading metrics per ad-set/inventory with Zena vs. benchmarks; spend pacing; secondary metrics.
• Pre-launch to-do checklist — the summary checklist from the Checklist reference.

What good looks like

• The budget calculator's CTR field is decorative (never used in its math) — do not gate the budget on CTR. CTR is only used later for the "pause after 1,000 impressions" launch guidance (benchmark-based).
• Keep revenue and ACV on the same basis (MRR with MRR, or ARR with ARR); convert if the user mixes them.
• Treat all outputs as directional planning figures, not precise forecasts — say so on the document.
• Design the document for light mode, self-contained (inline CSS, no external scripts or fonts, inline-SVG flowchart). Always deliver an exported PDF file as well as the HTML.
• One Campaign block per target audience; if the user has one audience, produce one Campaign block.