Build a Bottom of Funnel Keyword Strategy
By Sarthak Arora · From the SEO & Content collection · Updated July 2026
This prompt turns raw sales calls, customer pain points, paid ads data, and competitor gaps into a ranked keyword list that starts at the bottom of the funnel and works upward. It exists to fix the most common SEO failure: pouring effort into high volume top of funnel terms that never convert. You paste in what you know about the business, and you get back a prioritized list of buyer intent keywords, each scored for business impact and mapped to a page type the search results actually reward.
When to use this
- Organic traffic looks healthy but almost none of it turns into revenue or pipeline.
- You are about to hand a writer or content team a roadmap and want it aimed at money keywords first.
- Your top traffic page is a definition or 101 page ("what is a KPI") that no buyer would ever act on.
- You have access to sales call recordings, paid ads data, or a real customer base and want to convert that raw insight into search targets competitors have overlooked.
Fill in the variables
BUSINESS_AND_PRODUCT
The offer and the problem it solves, e.g. "identity theft protection sold to consumers who fear their data is on the dark web."
PRICE_AND_SALES_MOTION
E.g. "self serve, $12 per month, instant signup" vs "sales led, six figure annual contract, four month cycle." This shifts whether conversions happen on the page or much later.
SALES_CALL_NOTES
Paste real transcripts or verbatim pains ("what if employees find out we monitor them, how do we disclose it positively"). This is the input that produces keywords competitors miss.
PAID_ADS_DATA
The non brand terms that already convert in paid search, so organic mirrors proven demand.
COMPETITORS
Mix direct product competitors with content only sites; include any you believe you can outrank.
DIFFERENTIATOR
The objectively defensible difference, sourced from your internal expert if you have one.
AUTHORITY
Your rough domain strength versus each competitor, so the ranking accounts for what you can realistically win.
{{SERP_NOTES}} (optional): for terms you have already searched, note what ranks on page one today, e.g. "for 'employee monitoring software': two listicles, one vendor page, one forum thread." This lets the model mark those terms VERIFIED instead of guessing at the winning format.
The prompt
Full method. Works on any model.
You are a senior SEO strategist who has built revenue driven organic programs for software and ecommerce brands. You reject the volume first default: you start at the bottom of the funnel, mine primary sources most keyword tools never see, and rank targets by business impact, not search volume. Your output is a roadmap a content team can execute without further guessing. CONTEXT INTAKE Read the inputs below. {{SALES_CALL_NOTES}} and {{DIFFERENTIATOR}} matter most: if either is missing or thin, ask me for it before producing the list, one specific question at a time. If I confirm I cannot supply it, proceed, but label every keyword that depends on it HYPOTHESIS. Do not invent a business you were not told about. → {{BUSINESS_AND_PRODUCT}}: what you sell, to whom, and the exact problem it solves. → {{PRICE_AND_SALES_MOTION}}: price point, self serve vs sales led, sales cycle length. → {{SALES_CALL_NOTES}}: quotes, pain points, and objections from real discovery calls, support tickets, or customer interviews. Paste transcripts or phrases verbatim. → {{PAID_ADS_DATA}}: which paid search terms already convert (and which are just brand). → {{COMPETITORS}}: 3 to 6 direct and content competitors, plus any you can outrank. → {{DIFFERENTIATOR}}: what is objectively, defensibly different about your offer. → {{AUTHORITY}}: rough domain strength vs competitors (a third party score is fine). → {{SERP_NOTES}} (optional): for any candidate terms already searched, what ranks on page one today (page types, who owns them). METHOD (run in order) 1. SOURCE HIGH INTENT TERMS FROM PRIMARY DATA, not keyword tools first. Keyword tools sort by volume, which every competitor already sees, so they hand you no edge. a. Mine {{SALES_CALL_NOTES}}: extract the exact pains and phrases buyers use. Search the notes for seed phrases and read the language around each mention to surface ultra long tail queries in the customer's own words. Many terms that do not look bottom of funnel actually are; you only learn this from real calls. b. Mirror {{PAID_ADS_DATA}}: replicate the non brand terms that already convert in paid. Discard brand and competitor brand terms unless the goal is defensive. c. Run a content gap against {{COMPETITORS}}: list terms they rank for that a page like yours could win. Weight terms held by competitors weaker than you. 2. CLASSIFY EACH TERM BY BUYER INTENT. → Explicit commercial: "best [category] software", "[category] platform", "[X] vendor". → Comparison and alternatives: "[competitor] alternatives", "[A] vs [B]". These are the single highest converting type; searchers are near a decision. → High intent pain or how to with no commercial modifier that still convert ("warning signs of [problem]", "how to tell if [X]"). Do not overlook these. → Informational or definitional: 101 content with no buying intent. Flag to avoid. 3. VALIDATE INTENT AGAINST THE SEARCH RESULTS for every candidate. The top mistake is "I sell X so I should rank for X." What matters is who ranks now: if the page type winning the result does not match the page you can build, the term is unwinnable. → If you can browse live results, or {{SERP_NOTES}} covers the term, use that evidence and label the term VERIFIED. → Otherwise, predict the likely winning format from the term's wording and modifiers, label it UNVERIFIED, and never present a prediction as an observed result. Only target a term when the result rewards a format you can ship (product page, comparison post, collection page). Treat any listed difficulty score as a backlink read only; a low score can hide harder competition and a high score can hide a weak, beatable result. 4. SCORE BUSINESS IMPACT (High / Medium / Low / Zero) on two axes: (a) does the product directly solve the searcher's problem, and (b) how likely is the searcher to be in market and solution seeking. → HIGH: product solves it AND searcher is in market. Prioritize, but expect the most competition and noise. Example shape: "embedded analytics platform". → MEDIUM (the sweet spot): product could solve it, intent mixed, results blend informational and commercial. Less brutal competition. Example shape: a 2 to 3 word fathead term where results mix "what is" pages with "best of" pages. → LOW: partial fit, searcher unlikely to be buying. Mostly informational. → ZERO: no buying intent, pure definitions. AI answers strip this traffic. Avoid. 5. RANK AND SEQUENCE. Order by business impact first, then by how attainable the result is given {{AUTHORITY}} and {{DIFFERENTIATOR}}. Load the first 1 to 3 months almost entirely with bottom of funnel terms: they earn the most, are often lower difficulty, and early revenue keeps the program funded. Smaller players skip head terms owned by giants and go a layer or two deeper to attainable long tail variants. Do not spray many pages hoping one ranks; commit to the priority pages and drive them. OUTPUT FORMAT 1. A one paragraph read on where this business should focus and why. 2. A table of 20 to 40 keywords (fewer if the inputs are thin; say so), one row each, sorted by Priority, where Priority is a build order number (1 means build first) and Status is VERIFIED, UNVERIFIED, or HYPOTHESIS: | Priority | Keyword | Intent type | Winning page format | Business impact | Difficulty read | Status | Source | Why it converts | 3. A sequenced roadmap: which keywords to build in months 1 to 3, then months 4 to 6, with one line on why each phase is ordered that way. 4. The 3 to 5 highest priority terms with a one line justification each, plus any term you deliberately excluded and why. 5. A manual verification checklist: every UNVERIFIED or HYPOTHESIS row, with exactly what I should confirm in live search results before assigning it to a writer. SELF CHECK before returning → Verify: does every row carry an honest Status, with no predicted result format presented as an observed one? → Verify: is each keyword tied to a page format you can actually build and win? → Verify: does the months 1 to 3 phase skew almost entirely bottom of funnel? → Failure modes to avoid: ranking terms by volume; keeping definitional Zero impact terms; targeting head terms an incumbent has held for years; claiming a differentiator the inputs do not support ("we are cheaper" and "all in one" are not defensible).
For the most capable models. Goal and quality bar up front.
You are a senior SEO strategist who builds revenue driven organic programs. GOAL: turn my primary data into a bottom of funnel keyword roadmap a content team can execute without further guessing. Your first line of output is a one paragraph verdict on where this business should focus and why; supporting tables and detail follow. CONTEXT → {{BUSINESS_AND_PRODUCT}}, {{PRICE_AND_SALES_MOTION}}, {{DIFFERENTIATOR}}, {{AUTHORITY}} → {{SALES_CALL_NOTES}}: verbatim buyer pains, objections, and phrases from real calls, tickets, or interviews. → {{PAID_ADS_DATA}}: non brand paid terms that already convert. → {{COMPETITORS}}: direct and content competitors, flagging any you can outrank. → {{SERP_NOTES}} (optional): what ranks on page one today for any term already searched. {{SALES_CALL_NOTES}} and {{DIFFERENTIATOR}} carry the most weight. If either is missing or thin, ask one focused question before proceeding. PRINCIPLES → Start at the bottom of the funnel and mine primary sources first; keyword tools sort by volume, which every competitor already sees, so they give you no edge. → Classify each term by buyer intent. Comparison and alternatives terms convert highest. High intent pain and how to terms with no commercial modifier still convert; do not overlook them. Definitional 101 terms carry no buying intent; flag and exclude. → Validate intent against who ranks now, not "I sell X so I should rank for X." Target a term only when the winning result rewards a format you can ship. Treat any difficulty score as a backlink read only. → Score business impact on two axes: does the product solve the searcher's problem, and is the searcher in market. Head terms win most but face the fiercest competition; the two to three word fathead terms with mixed results are usually the sweet spot. → Rank by business impact, then attainability given {{AUTHORITY}} and {{DIFFERENTIATOR}}. Load months 1 to 3 almost entirely with bottom of funnel terms so early revenue funds the program. Commit to priority pages rather than spraying many and hoping. QUALITY BAR → 20 to 40 keywords (fewer if inputs are thin; say so), each tied to buyer intent, a business impact tier, and the page format its results reward. No ranking by volume. → Several terms trace to language pulled straight from {{SALES_CALL_NOTES}}. → Every row carries an honest Status: VERIFIED when confirmed from real results, UNVERIFIED for predicted formats, HYPOTHESIS when a required input was missing. → A sequenced roadmap (months 1 to 3, then 4 to 6) plus a checklist of exactly what to confirm in live search for every UNVERIFIED or HYPOTHESIS row before briefing a writer. BOUNDARIES → Do not invent a business, statistics, or SERP results you were not given. → Never present a predicted format as an observed one. → Do not pad with generic SEO advice or claim a differentiator the inputs do not support ("cheaper" and "all in one" are not defensible). → If a required input is missing, ask one focused question instead of guessing.
Five lines. Speed over rigor.
Act as a revenue focused SEO strategist. From {{SALES_CALL_NOTES}}, {{PAID_ADS_DATA}}, and {{COMPETITORS}}, give me 15 to 25 bottom of funnel keywords (comparison, alternatives, high intent pain and how to), each with buyer intent, business impact (High/Medium/Low), and the page format its results reward. Rank by business impact, not volume, and skip pure definitional terms. Mark any format you predicted rather than confirmed as UNVERIFIED.
Want all 120 prompts in one workspace?
Every prompt in this library, organized by task. Free.
What good output looks like
- Every keyword is tied to buyer intent and a business impact tier, never ranked by raw search volume.
- At least a few terms trace back to language pulled from real sales calls or customer research, phrased the way a buyer actually types.
Show 4 more quality checks
- Each term names the page format the search results reward, and terms whose results demand a format you cannot build are explicitly excluded.
- Every row carries an honest Status: formats confirmed from real results are marked VERIFIED, predictions are marked UNVERIFIED, and the output ends with a checklist of exactly what to confirm in live results before briefing a writer.
- The roadmap sequences months 1 to 3 almost entirely bottom of funnel (comparison, alternatives, high intent how to), with head terms sequenced later or deliberately skipped.
- Missing inputs surface as HYPOTHESIS rows to validate, not dressed up as confident findings.
Related prompts
- Decide Whether SEO Is Worth It
Run this first if you are not sure search is even the right channel for this business.
- Write a SERP Driven Content Brief
Take a prioritized keyword from this list and turn it into a brief a writer can execute.
- Design Topic Clusters and Internal Links
Group the winning keywords into clusters and internal links so the money pages rank faster.
- Structure a Google Ads Account
Put the same buyer intent keywords to work in paid search while SEO compounds.
Free to use and share. If you republish a prompt, link back to this library.