SEO Discovery and Goal Setting
Run SEO discovery that connects business context, customer economics, constraints, scope, and measurable goals before audits or keyword research begin.
SEO discovery is the phase that turns “improve SEO” into a commercial assignment. It establishes what the business sells, who buys, what creates value, how progress is proved, and what can ship. Without that context, a team can perform an excellent technical audit, keyword study, or content program and still optimize efficiently toward the wrong outcome.
Phase: P0, Stage A — Understand. Timebox: three to five working days for one business, including one 90-minute decision workshop and focused follow-ups. Owner: the SEO or growth lead, accountable to a business sponsor who can approve economics and scope. Sales, product, finance, analytics, engineering, legal, and regional owners contribute where their decisions affect the plan.
Why this phase exists and why it comes first
Every later choice is downstream of business context. Post-type priorities depend on whether the company is a recurring software product, a catalog, a two-sided platform, a local operator, an audience business, or an expert service. Keyword and prompt choices depend on the ideal customer profile (ICP), meaning the type of account or buyer most likely to receive value and become a good customer. Page priorities depend on product margin, sales capacity, market readiness, and the value of the conversion a page can influence.
P0 has no earlier analytical phase. It consumes the project mandate, access to decision-makers, and the records the business already uses to run sales and marketing. It produces the contract that Phase P1, access, tracking, and data sources, uses to decide which systems must be connected and which events must be measurable.
Starting with a technical audit reverses that dependency. The audit can show what is broken, but not whether fixing it supports the valuable product, buyer, or region. Starting with keyword volume has the same defect: demand can be real while the resulting customers are unprofitable, outside the service area, or unlikely to pass legal and sales qualification. Skipping discovery is therefore the most expensive sequencing mistake because its error propagates into every brief, fix, measurement, and report that follows.
Inputs and outputs
An input is not complete because someone shared a deck. It must be current enough and specific enough to support a decision. An output is not complete because the workshop happened. It must record the decision in a form the next phase can cite.
P0 inputs and outputs
| Direction | Item | Why it is needed | Acceptance condition |
|---|---|---|---|
| Input | Business and product overview | Connects visibility work to what is sold and how money is made. | Lists active offers, revenue model, strategic products, margins or approved proxies, markets, and seasonality. |
| Input | Customer and sales evidence | Separates the actual buyer from a generic audience description. | Names ICP, buying committee roles, qualification rules, sales-cycle range, objections, and conversion stages. |
| Input | Performance and financial records | Provides a baseline in the numbers the business already trusts. | Includes source, owner, definition, date range, and current value for candidate KPIs. |
| Input | Operating constraints | Prevents the plan from assuming capacity or permissions that do not exist. | Covers development, content management system (CMS), design, legal, brand, localization, data, and publishing throughput. |
| Output | Business-type classification | Sets the initial page and post-type priorities. | Selects one primary classification, records any secondary model, and cites supporting evidence. |
| Output | Goal and key performance indicator (KPI) contract | Gives every later report one commercial interpretation. | Names one primary outcome and KPI with formula, baseline, source, window, owner, and target or decision rule. |
| Output | Constraint and assumption register | Makes feasibility and uncertainty visible before planning. | Every item has impact, owner, status, and resolution or review date. |
| Output | Scope agreement | Protects the project when new requests appear later. | Lists included and excluded work, deliverables, capacity, dependencies, and the change-control rule. |
These outputs are the contract with P1. If the primary KPI is qualified demos, P1 must verify the form, CRM stage, and source attribution. If the outcome is contribution from organic orders, P1 must connect search and commerce data and preserve the margin definition. A generic request for “analytics access” is not an adequate handoff.
Ready to Monitor Your AI Visibility?
Track how AI chatbots mention your brand across ChatGPT, Perplexity, and other platforms.
The discovery checklist
Each item below includes the action, its reason, the method, the working tool, and an observable done-when condition. Store evidence beside the decision; memory and meeting notes are not a control system.
1. Classify the business model
What to do: Select the primary model from the business-type strategy hub : ecommerce , SaaS , local service , marketplace , media, publisher, or affiliate , or B2B services . Record a secondary model only when it changes priorities.
Why it matters: The model predicts the useful starting page set. A catalog needs category and product coverage; SaaS often needs use-case and comparison demand; a marketplace must serve both supply and demand. Treating all three as “content sites” hides their different growth mechanics.
How to do it: Ask how the company earns money, what unit is sold, who supplies it, whether revenue repeats, and which transaction or contract ends the journey. Test the answer against real revenue, not the homepage tagline.
Tool: Business model interview, product catalog, revenue report, and the six business-type playbooks.
Done when: One primary type is named, the evidence is cited, any hybrid is explained, and the sponsor agrees that its initial post-type priorities fit the growth model.
2. Map products, economics, and seasonality
What to do: List offers and product families with revenue model, price or average order value, gross margin, contribution margin, refund or churn exposure, strategic priority, capacity, and seasonal peaks. Gross margin is revenue after direct production cost; contribution margin also subtracts the variable costs required to make the sale.
Why it matters: Revenue is not equally valuable. A high-volume category with thin margin, constrained inventory, or poor retention may deserve less acquisition effort than a smaller offer the business can fulfill profitably.
How to do it: Reconcile the product leader’s priorities with finance or commerce records. Cover products accounting for at least 80% of trailing revenue plus every explicitly strategic launch, then document the long tail as a later decision. Use a full prior year when seasonality exists; label a shorter period as provisional.
Tool: Product catalog, finance report, order system, retention report, and AmICited Economics for connected ecommerce data.
Done when: Every in-scope offer has an economic value or approved proxy, a capacity flag, a seasonality note, and a named priority. Unknown margin is visible rather than treated as zero.
3. Define the ICP and buying committee
What to do: Describe the ICP and each buying-committee role. A buying committee is the set of people who influence, approve, use, finance, or block a purchase. For consumer sales, replace committee roles with shopper, payer, recipient, and other relevant influencers.
Why it matters: Different roles ask different questions. A user may search for workflow fit while security checks controls, finance evaluates payback, and an executive asks about business risk. One vague persona cannot support all four decisions.
How to do it: Use won and lost deals, sales calls, support conversations, reviews, CRM fields, and customer interviews. Record firmographic or situational fit, trigger, job to be done, objections, disqualifiers, and the role that owns the final decision.
Tool: Customer relationship management (CRM) system, call recordings, win-loss notes, review corpus, and a structured interview worksheet.
Done when: The primary ICP is narrow enough for sales to accept or reject an example account, every material role has questions and objections, and disqualifiers are explicit.
4. Define the conversion and its value
What to do: Name the primary conversion, meaning the observable action that advances a qualified buyer toward value, and estimate what it is worth. Record secondary conversions separately.
Why it matters: A form submission, trial, order, phone call, subscription, and store visit have different economic meanings. Counting them as interchangeable “conversions” makes prioritization and reporting unreliable.
How to do it: Trace the path from visit to sale and from sale to retained value. For lead generation, combine qualification rate, close rate, and expected margin or lifetime value, the value expected across the customer relationship. For commerce, start with contribution rather than gross revenue when cost data permits. If value is uncertain, use labeled low, base, and high cases.
Tool: Analytics event dictionary, CRM funnel, commerce system, finance model, and attribution records.
Done when: The conversion has an exact event or stage definition, owner, source, current rate or count, and monetary value or documented range. Secondary actions cannot silently substitute for it.
5. Separate the goal, KPI, and vanity metrics
What to do: Write one goal, one primary key performance indicator (KPI), and the supporting indicators. A goal is the business outcome to create. A KPI is the quantified measure used to judge progress toward it. A vanity metric is a number that can rise without demonstrating meaningful progress.
Why it matters: “More traffic” confuses exposure with value. Organic sessions can grow through irrelevant queries while qualified pipeline, orders, or margin stays flat. Traffic may diagnose the path, but it does not define success.
How to do it: Write the goal in the client’s own numbers: “increase monthly contribution from non-brand organic orders” or “increase sales-qualified demos attributed to organic landing pages.” Define formula, scope, window, source, baseline, target, and owner. Then label impressions, rankings, clicks, mentions, and raw leads as supporting metrics unless one is genuinely the contracted outcome.
Tool: Discovery workshop, KPI dictionary, analytics, CRM, finance, and executive operating report.
Done when: One primary outcome and one KPI fit on a single line; the baseline can be reproduced from the named source; and every supporting metric has a stated relationship to the outcome.
6. Reconcile business competitors with search and AI competitors
What to do: Create two lists. Client-defined competitors sell a comparable solution or compete for budget. Search and AI competitors occupy results or answers for the same customer questions, even when they monetize differently.
Why it matters: The client may name three familiar brands while search results favor publishers, directories, marketplaces, communities, or a new entrant. Using only the client list misses the pages actually winning discovery; using only search results misses commercial substitutes sales encounters.
How to do it: Collect the client list, then review at least 20 representative non-brand queries and prompts across awareness, comparison, and action intent. Record recurring domains and brands, page type, appearance count, and why each competes. Use AmICited Competitor Analysis to see brands appearing beside the company in tracked AI answers.
Tool: Stakeholder interview, search result review, prompt set, and https://app.amicited.com/competitors.
Done when: Both lists exist, overlap and differences are explained, evidence covers all three intent stages, and the working benchmark set has an owner and review date.
7. Establish geography, language, and market readiness
What to do: Define countries, service areas, languages, currencies, legal entities, and localization expectations. Separate where the company can sell from where it merely receives traffic.
Why it matters: A ranking in an unsupported country or language creates noise, not useful growth. Market-level margin, fulfillment, regulation, and sales coverage can change the value of identical demand.
How to do it: Compare commercial priorities with orders, qualified pipeline, service coverage, shipping rules, language ownership, and contribution by market. Connected ecommerce teams can use AmICited Geography to inspect revenue and contribution by country and city.
Tool: CRM or commerce geography, fulfillment rules, regional P&L, localization inventory, and https://app.amicited.com/reports/geography.
Done when: Every in-scope market has a language, offer, fulfillment or sales owner, conversion path, and economic rationale; excluded markets are recorded explicitly.
8. Capture constraints and agree scope
What to do: Record development capacity, CMS limitations, legal or regulatory review, brand restrictions, publishing throughput, design capacity, data gaps, and approval lead times. Convert them into included work, excluded work, dependencies, and change control.
Why it matters: Constraints change sequencing. A six-week legal gate favors early review; a CMS that cannot produce indexable category copy changes the implementation path; one developer-day per sprint rules out a plan built around weekly template work.
How to do it: Ask each owner what they can approve or ship per week, what commonly blocks work, and the longest lead time. Plan committed production to no more than 80% of confirmed throughput so revisions and incidents have capacity. Define a scope-change trigger for a new market, language, product family, deliverable, review gate, or more than 10% additional planned effort.
Tool: Capacity plan, CMS capability review, approval map, risk register, and written statement of work.
Done when: Every constraint has impact, owner, and lead time; in-scope and out-of-scope work are listed; and the sponsor has accepted the change-request path before delivery starts.
Tools in AmICited
AmICited supplements business interviews; it does not replace them. Use product data only where a connected report answers a discovery question, and record the report window beside the decision.
- Open
https://app.amicited.com/competitorsand use Competitor Analysis to compare stakeholder-named rivals with brands detected in AI answers. Export or record the evidence date and prompt scope.
- For ecommerce, open
https://app.amicited.com/reports/economicsand use Economics to distinguish revenue from contribution and identify missing cost inputs before assigning product value.
- Open
https://app.amicited.com/reports/geographyand use Geography to test whether proposed markets have orders, revenue, and contribution evidence. Treat a market below the report’s 30-order margin floor as unmeasured, not unprofitable.
If a required source is not connected or its definitions do not match the client’s operating report, record the gap for P1. Do not manufacture a baseline by combining incompatible windows or definitions.
Stay Updated on AI Visibility Trends
Get the latest insights on AI mentions, brand monitoring, and optimization strategies.
Decision rules
Thresholds here are operating rules for this process, not claims about universal business performance. They force ambiguity into a visible decision.
P0 decision rules
| Finding | Threshold | Decision | Evidence required to pass |
|---|---|---|---|
| Primary outcome | Zero or more than one outcome marked primary | Hold prioritization | Sponsor selects exactly one; other outcomes become secondary or receive separate scope. |
| KPI definition | Any of formula, source, window, owner, baseline, or scope is missing | Reject the KPI contract | A second person can reproduce the baseline from the named source. |
| Baseline window | Shorter than 28 days or one full sales cycle, whichever is longer | Mark provisional | Use a longer comparable window or document why no stable history exists. |
| Product coverage | Less than 80% of trailing revenue covered, or a strategic launch omitted | Extend the product map | Covered set reaches the threshold and includes every sponsor-named strategic offer. |
| Competitive evidence | Fewer than 20 representative non-brand queries and prompts or one intent stage missing | Keep benchmark set provisional | Evidence spans awareness, comparison, and action intent with date and market recorded. |
| Market economics | Fewer than 30 orders in AmICited Geography | Do not compare margin | Use revenue or order evidence only and label margin unmeasured until the floor is met. |
| Delivery capacity | Committed work exceeds 80% of confirmed throughput | Reduce or phase scope | Named owners confirm capacity with revision allowance preserved. |
| Scope change | New market, language, product family, deliverable, approval gate, or over 10% added effort | Open change control | Impact on time, cost, KPI, dependencies, and displaced work is approved in writing. |
An unresolved unknown is not automatically a failure. It becomes a failure when it can change priority or measurement and has no owner, resolution date, or provisional rule.
Deliverable: the discovery document
Hand over one versioned discovery document, not a slide deck plus scattered notes. Markdown, a shared document, or a structured project record is acceptable if it contains the same contract:
# P0 Discovery Document
Decision owner: [business sponsor]
Delivery owner: [SEO/growth lead]
Approved on: [date]
## Business context
- Primary business type and evidence:
- Secondary model, if material:
- In-scope products and economic value/proxy:
- ICP, buying roles, sales-cycle range, seasonality:
- Markets and languages:
## Goal and measurement contract
- Primary business outcome:
- Primary KPI and formula:
- Baseline value, period, source, and extraction date:
- Target or decision rule:
- Supporting metrics and their purpose:
## Competitor sets
- Client-defined competitors:
- Search and AI competitors:
- Evidence set, market, date, and review owner:
## Constraints and assumptions
- Constraint | impact | owner | lead time | status/review date
- Assumption | consequence if false | owner | resolution date
## Scope
- Included:
- Explicitly excluded:
- Deliverables and throughput:
- Dependencies:
- Change-control trigger and approver:
Attach source exports or links, but keep the approved decisions in this record. The sponsor approves business outcome, economics, market, and scope. The delivery lead confirms that the work is feasible. Analytics, finance, sales, and product owners verify only the definitions and evidence they own.
What goes wrong
Accepting “more traffic” as the goal. This makes a channel metric the destination. Replace it with a business outcome and preserve traffic as a diagnostic measure with an explicit role.
Discovering a review gate in month three. Legal, medical, financial, security, brand, or localization review changes lead time and throughput. Identify the approver, evidence standard, and normal turnaround before the first brief enters production.
Assuming the business model from the website. A company may present as SaaS while services produce most margin, or operate a marketplace while earning through media. Classify from revenue mechanics and customer behavior, then reconcile the public story.
Using revenue without margin or capacity. Revenue can prioritize an offer the business cannot fulfill or barely profits from. Use contribution where available and label proxies honestly.
Treating the client’s competitors as the complete benchmark. Commercial rivals and result-page rivals answer different questions. Keep both lists and explain the overlap.
Creating a scope with no change rule. The out-of-scope request in week six is not resolved by goodwill. Record whether it is rejected, swapped for committed work, or estimated as a separate change, and name the approver.
Hiding unknowns inside polished language. “Value to be confirmed” is not a plan. Record the estimate range, consequence, owner, and deadline or stop the downstream decision it affects.
Handoff to P1: access, tracking, and data sources
P1 receives the approved discovery document and converts its measurement contract into an access and instrumentation plan. The handoff must identify the primary KPI, conversion definition, baseline source, relevant systems, markets, languages, comparison window, and the owners authorized to grant access.
P1 should be able to answer five questions without reopening discovery:
- Which business outcome is primary?
- Which event, order, or CRM stage represents it?
- Which source contains the baseline and who owns that source?
- Which segments, products, markets, and languages are in scope?
- Which constraints affect connection, consent, implementation, or review?
If P1 finds that the event does not exist, a source uses a conflicting definition, or access cannot be granted, it returns a named measurement gap to the sponsor. It does not silently replace the KPI with sessions, impressions, or another convenient metric. Every later phase cites the same approved discovery document until a material change reopens P0.
FAQ
Frequently asked questions
How long should SEO discovery take?
Is more organic traffic a valid SEO goal?
What if the client cannot provide a conversion value?
Who approves the discovery document?
When should discovery be reopened?
More tutorials in this section
Ready to put it into practice?
Free check · 7-day trial · no credit card