# FOCUSADS methodology > The six planning stages, the eight weighted scoring factors and the hard exclusions that override any score. Source page: https://focusads.si/en-gb/methodology Updated: 2026-10-24 We start from the result your finance team will count, then work back to the buyer, the media and the evidence needed before money moves. FOCUSADS Superintelligence starts every plan from the business result: what must change, and what the business can afford to pay for that change. The answer sets which buyers matter, which channels earn a role in your market, what has to be measured before launch and when budget moves. The same six stages run on every account, from a one-market pilot to a programme across several countries. ## Three rules above every stage - The business result outranks the platform metric: A platform conversion is a signal. Revenue, a qualified opportunity or a kept appointment is a result. When the two disagree, the result wins and the optimisation target changes. - Every channel gets a job description: Capture demand, create it, convert it, recover it or reach people away from screens. A channel without a stated role cannot be judged, so nobody can decide on evidence whether to scale it or cut it. - Measurement is agreed before the money: What counts as success, which system holds the truth and how long a test runs are written down before the first platform invoice arrives. ## Six stages, in the order we run them ### 1. Define success and what it may cost What must change in the business, and what can it afford to pay per result? We fix one primary outcome (revenue, qualified leads, booked appointments, pipeline value, dealer applications, retention, or awareness when awareness really is the goal) and write a definition that sales and finance sign. Then the economics: monthly media budget, average order or contract value, margin, lead-to-sale rate, sales cycle and how many leads your team can actually follow up. From these we derive the most you can pay for a customer and for a qualified lead. If the numbers show paid acquisition cannot pay back, the plan says so before anyone spends. What gets decided: - Primary outcome, its written definition and who confirms it happened (CRM, store platform or finance) - Permissible cost per customer = contribution per customer × share available for acquisition - Permissible cost per qualified lead = permissible cost per customer × qualified-lead close rate Evidence kept: Signed outcome definition, economic inputs, assumptions and formula version, stored with the plan. If this stage is skipped: Campaigns optimise to whatever the platform counts as a conversion, and a "good" cost per lead is declared without knowing what a lead is worth. ### 2. Identify the buyer and the trigger Who exactly has to act, and what makes them act now? A Goal Segment names a buyer and a buying trigger: a fleet manager whose lease contracts come up for renewal, a first-time buyer comparing new-build flats, a distributor looking for a second supplier. The reference model covers 17 industries, 169 sub-industries and 68 Goal Segments, each with its usual journey, sales cycle and compliance limits. We agree with your sales team what makes a lead real, so the qualification rule exists before the first lead does. What gets decided: - Primary Goal Segment and its buying trigger - Qualification rule: what turns an enquiry into a qualified lead - Targeting the category or market does not allow (for example health, housing or credit restrictions) Evidence kept: Segment definition and qualification criteria agreed with sales, saved with the plan. If this stage is skipped: Broad targeting fills the funnel with people who will never buy, and the bidding algorithm learns from them. ### 3. Select media for this buyer in this market Which channels earn a place, and what is each one for? Every candidate channel, online or offline, gets a priority score from 0 to 100 built from eight weighted factors (listed below). Before any score counts, hard exclusions run: a channel that is unavailable or restricted in your market, prohibited for your category, missing a prerequisite such as a product feed or an app, or whose workable minimum does not fit your budget is removed, whatever its score. Local sellers come from our market media registry, where a platform is only listed as available after we have checked the seller’s own documentation for that country. The budget is then split across a small number of channels so each gets enough data to learn. What gets decided: - Primary, supporting and test channels, each with a stated role - Why each channel ranked where it did, and which excluded channels were removed and why - Media split that adds up to your monthly media budget, with a per-channel minimum Evidence kept: Scored channel table with stored factor points, reason codes, exclusions and the scoring version. If this stage is skipped: The budget is spread across six platforms, none of which collects enough data to learn. ### 4. Validate the landing experience and tracking Can the click become a customer, and will we see it happen? For lead and sales goals the landing page and the tracking are working parts of the plan. We decide whether your existing page can carry the campaign, needs fixes, or whether a dedicated page is required per market or segment, and we price that decision. Tracking scope covers GA4, tag management, platform tags, consent mode, enhanced or server-side conversions where they apply, and the import of CRM outcomes. Test conversions are recorded and consent behaviour is checked per market before launch. What gets decided: - Landing-page decision with its cost - Tracking scope with its cost - Consent and privacy configuration for each target market Evidence kept: Recorded test conversions and a consent check, attached to the launch gate. If this stage is skipped: Paid traffic lands on a generic page, and reported conversions cannot be reconciled with sales. ### 5. Launch and verify What will we look at, when, and what counts as working? Before spend starts we agree the source of truth for each metric, the attribution windows, the review cadence and the minimum data a channel needs before anyone judges it. Launch QA covers links, forms, events, budgets, naming, platform policy and consent. In the first weeks we verify that the numbers in the ad platform, the analytics tool and your CRM tell the same story, and we fix the gaps before optimising anything. What gets decided: - Metric definitions and their sources - Review cadence and decision thresholds - Go / no-go checklist, signed off Evidence kept: Completed launch checklist with evidence attached, and the first reconciliation note. If this stage is skipped: A channel is paused in week two before it has enough data, or keeps running long after it should have stopped. ### 6. Reallocate on evidence Where should the next pound go? Optimisation follows the outcome: offline conversion imports, lead-quality feedback from sales, search term and placement reviews, creative fatigue checks and landing-page tests. Budget moves towards channels and segments whose cost per result sits inside the permissible range with enough volume to trust it. Channels that miss it for consecutive review windows are restructured or stopped. New markets and channels enter as tests with a fixed budget and a stop rule. You approve every reallocation before budgets change. What gets decided: - Scale, hold, restructure or stop, per channel - Test budget and stop rule for any new channel or market - Updated plan and costs, sent for your approval Evidence kept: Change log and a reallocation proposal you approved, kept with the account. If this stage is skipped: Budgets grow with the calendar instead of with evidence, and cheap leads that sales rejects keep getting more money. ## How the priority score is built Scoring model v2.1, updated 2026-10-26. Each factor is rated 0 to 100 for the channel and your answers, then multiplied by its weight. The weighted points add up to the priority score. The weights below are the ones the planner uses today; they change only with a new scoring version. | Factor | Weight | What it asks | | --- | --- | --- | | Goal and commercial fit | 25 | Does the channel’s strongest format serve the result you chose, and can that result be tied to your economics? | | Industry and buyer intent | 20 | Are your buyers present here, and are they looking for what you sell or do they need to be persuaded first? | | Verified local reach | 15 | Is the product confirmed as available in your target markets from an official source? | | Budget viability | 15 | Does its workable minimum fit what is left after setup, landing page and tracking are paid for? | | Measurement readiness | 10 | Can its results be measured with the tracking you have, or does it need a lift test or CRM data? | | Creative and operations | 5 | Do you have the creative, feed or production it needs, and can your team support it? | | Timing and sales cycle | 5 | Does it suit your sales cycle and the time you have? | | Incremental reach | 5 | Does it reach people your current channels do not, or does it overlap them? | The number of channels is capped by budget so each one can learn: two below 3,000 USD of monthly media, three below 8,000, four below 20,000, five below 50,000, six above that. ## Hard exclusions override any score A channel that scores 92 is still removed if one of these applies. The plan shows every removed channel, the reason and what to do instead. - Excluded: FOCUSADS does not run campaigns in this market until a verified local route exists. - Excluded: the platform does not sell this ad product in your target markets. - Excluded: the platform restricts advertising in your target markets. - Excluded: availability in your target markets is not confirmed. - Excluded: FOCUSADS does not plan paid media for this product category. - Excluded: advertising this category to consumers is not permitted in your target markets. - Excluded: requires a live iOS app. - Excluded: requires a product catalog, and your business does not sell catalog products. - Excluded: your products must be listed on the marketplace first. - Not priced: no verified local media partner in our records yet, so a strategist plans it with you. - Excluded: its workable minimum is above your permitted media budget. - Excluded: it would produce more leads than your team can follow up. - Excluded: it reaches under 5% of searches in your target markets. - Excluded: its ads reach under 5% of internet users in your target markets. ## Human review and the audit trail The planner produces a recommendation, not a booking. A strategist reviews it with you before anything is bought. When the agreed plan differs from the computed ranking, the change and its reason are written into the approved plan. Every plan stores its inputs, factor points, exclusions, engine version and scoring version, so a plan from today can still be explained next year. In OPS, changes to work items, approvals and media research records are written to an audit log with who changed what and when. In delivery: a structured override field on each channel line, so strategist changes appear in the client dashboard next to the original score. ## How we read numbers Three tiers, always reported separately so nobody mistakes a signal for a result. - Business results: Revenue, closed deals, kept appointments, pipeline value, approved dealer applications (source: CRM, store platform or finance) - Qualified outcomes: Qualified leads, sales-accepted leads, booked meetings, add-to-basket from the target segment (source: CRM stages and the agreed qualification rules) - Platform signals: Platform conversions, CTR, CPC, CPM, view rate, frequency (source: The ad platforms’ own reporting, shown as diagnostics) ## The planning formula Plan = f(Market, Industry, Goal, Goal Segment, Economics, Maturity, Constraints) - All-in acquisition cost = media + attributable service fee + share of setup + other acquisition costs. - A permissible cost per qualified lead holds only for the lead definition and the cohort used to measure the close rate. - Every market-platform pair starts as a research candidate. It becomes eligible for a client only after account, market, format and policy checks are recorded. ## What we will not do - Promise a ROAS, lead volume or revenue figure before seeing your data. - Infer competitors’ ROAS or acquisition cost from public ad libraries. - Use sensitive health or protected-class targeting, or work around platform policy. - Run your advertising from accounts we own. You keep ownership; we work through access you can revoke. ## Questions about the methodology ### Does FOCUSADS Superintelligence use machine learning to score channels? Not for the score. The scoring model is a fixed rule set with eight published weights and stored factor points, so you can see why a channel scored 78 and not 90, and the same answers always give the same score under the same model version. Judgement comes in at two points: the strategist who reviews and signs the plan, and the tests after launch that show what a channel added. ### Can a strategist change the recommendation? Yes, during the plan review and only with you. The approved plan records the change and the reason. A structured override field that shows both numbers in the dashboard is in delivery. ### What if we do not know our margin or close rate? The plan still runs, with lower data confidence, and the report says so. Collecting those two numbers is usually the first client action, because without them nobody can judge whether a cost per lead is good. ### Do you plan offline media the same way? Yes. Outdoor, radio, TV, print and events get a role, a budget line and a measurement design such as geo tests, unique URLs or QR codes, call tracking where it is installed, and brand-search lift. Offline inventory is priced only when a local media owner is verified in our records; otherwise a strategist scopes it with you. ### How long before you scale a channel? When it has enough volume to trust the cost per result and stays inside the permissible range for at least one full review window. Those thresholds are agreed at launch, not chosen afterwards.