targetbay-email-sms

Use when designing a test, deciding whether a test is worth running, or reading a completed test's result — sizing variants, setting the decision threshold and duration in advance, choosing whether the design needs a holdout that receives nothing, and judging whether a difference is real or noise. Covers any dimension — offer, audience, product, timing, channel and content — not just subject lines. Use campaign-optimization when the question is what is wrong; this skill runs the experiment that answers it.

Use when deciding how much of a store's customer-facing writing a model may produce unattended — which automated message types may ship generated wording with no human sign-off, which must always be approved by a person before they reach a recipient, what that reviewer checks and in what order, whether a model may pick an incentive and from which pre-approved set, and what an unattended send does when the generator is unavailable or returns something unusable. Answers "can we let it write our newsletter?", "do we have to approve every generated email?" and "what stops an unattended send saying something untrue about our products?". Use content-optimization when one specific message underperforms and its wording needs rewriting rather than governing.

Use when average order value is the objective and the lever is not yet chosen — ranking thresholds, bundling, merchandising, tier moves and cross-category attachment against each other by revenue effect and margin cost. Owns free-shipping and gift threshold design. Answers "raise our average order value" and "where is our basket losing value?". Use upsell once the chosen lever is an upward move onto a premium tier, larger size or subscription, and cross-sell once it is attachment in another category.

Use when deciding who to target, which segment or list to send to, or whether an audience should be split into separate segments. Determines whether the answer is one audience or several, ranks them by expected value, defines who must be excluded, and confirms each segment actually exists and is large enough to be worth sending to. Answers "who should we target?", "which list?" and "should we segment this?". Other skills delegate targeting here, so segment definitions stay in one place.

Use when designing or restructuring an automation journey — a welcome series, post-purchase flow, abandoned cart or browse journey, replenishment, win-back or VIP track. Decides how many separate automations one objective needs, whether to split a journey by audience, product, category, order value or price band, how many nodes it should have, when each step fires and which channel carries it. Answers "should this be one automation or several?" and "should we split this flow?". Use automation-strategy when the question is which automations the store should have at all, and automation-optimization when an existing journey needs tuning rather than reshaping.

Use when a live automation is underperforming and needs tuning rather than redesign — a welcome series that is not converting, an abandoned cart flow whose revenue is down, a drip sequence with falling completion, or a journey that used to work and has declined. Finds which node loses people, whether the timing is wrong, whether the wrong contacts are entering, and whether the journey has decayed through expired offers, dead links or discontinued products. Use automation-architect instead when the journey's shape itself is wrong.

Use when part of a journey cannot live inside the marketing platform and the question is where each step should run — a trigger the platform does not emit, a wait longer than it supports, data it does not hold, a human approval step, an enrichment call, or a sync with a CRM, a spreadsheet or an advertising audience. Decides what stays inside the platform, what moves to an external workflow tool, and what the store gives up by moving it — suppression enforcement, frequency counting, attribution and reporting. Answers "should we build this inside the platform or outside it?", "how do we connect this to our other systems?" and "what triggers this journey?". Use automation-architect to design the journey itself and automation-strategy to decide which journeys should exist at all.

Use when a store wants to know which ready-made automation recipes to adopt from a published library and in what order — picking from a finite catalogue of trigger-by-trigger recipes instead of designing one from scratch, ranking each recipe by what this store can actually support today, naming the data, consent and integration prerequisite every recipe has before it can run, and saying plainly when none of them fit. Answers "which recipes should we use?", "is there a ready-made recipe for this?" and "what can we adopt with what we already have?". Use automation-strategy when the portfolio should be derived from this store's own data rather than picked from a library, and automation-architect to design a single journey in detail once one is chosen.

Use when the question is which automations a store should have at all — auditing the whole automation portfolio, finding lifecycle stages with no coverage, deciding which existing journeys should be split, consolidated, extended or retired, and sequencing them by expected value. Use opportunity-discovery instead when no one has said the problem is automation. Use this before automation-architect, which designs the topology of a single objective once this skill has decided the objective is worth building.

Use when a campaign and the automations already running would both reach the same customers at the same time, when two automations fire for the same contact, or when a seasonal push overlaps a lifecycle message or the same moment is planned on both email and SMS. Decides which send proceeds, which is delayed, which is suppressed for the overlap and which switches channel. Answers "our campaign overlaps with automations that are running, what do we do?", "two automations both fire for the same customers" and "which message wins?". Use marketing-calendar to sequence a plan before any of this arises, and channel-optimization for channel assignment absent contention.

Use when a campaign, newsletter or email blast underperformed and needs diagnosis — falling engagement, declining clicks, weak revenue, rising unsubscribes, or a recurring send that keeps getting worse. Works backwards from revenue and conversion to find the actual failure point — audience, offer, timing, channel or content — then forms a hypothesis and proposes a testable change. Also use to understand why a campaign did well. Never optimises on open rate alone.

Use when deciding how email and SMS should divide the work for an objective — which messages belong on which channel, which segments are reached where, how a paired email and SMS are ordered and how far apart, what each channel's version of the message actually says, and whether SMS is worth its cost for a given moment. Answers "should this be email, SMS, or both?", "what should we send on SMS?" and "how do we coordinate the two channels?". Use campaign-conflict-resolver when two sends already contend for the same contact, and audience-discovery when the question is who to reach rather than where.

Use when setting the standing policy for who a store may contact and how much — whether people must confirm before anything is sent to them, what to do with contacts imported or inherited with no provable consent record, and separately how many messages a segment should receive in a period across email and SMS together before contact becomes a cost. Answers "should we use double opt-in?", "can we mail this list we acquired?" and "how often is too often to email our customers?". Use list-hygiene when the list is consented and the problem is bounces, and campaign-conflict-resolver when specific sends already breach the ceiling this skill sets.

Use when the message itself is the problem or is being written — subject lines and preheaders that under-deliver, body copy that does not convert, an offer stated ambiguously, or content that no longer matches the audience it goes to. Also use to decide what a message should contain and in what order — the lead, the proof, the product blocks, the single call to action. Produces content direction and testable variants, never finished creative. Answers "rewrite this subject line", "what should this email actually say?" and "why is nobody clicking?". Use campaign-optimization first if content is not yet established as the failure point.

Use when the objective is getting existing customers to buy from a different product or category than they already have — expanding category breadth, promoting complementary products after a purchase, or moving single-category buyers into a second category. Use upsell when the goal is a higher-value version of the same thing.

Use when the whole customer journey needs designing or reviewing rather than one stage — mapping where customers actually sit, deriving the store's own stage boundaries, finding which transitions have no coverage, and sequencing the lifecycle programme by where value leaks. Use customer-retention or customer-winback when one stage is already the agreed focus.

Use when the objective is keeping existing customers active — improving repeat purchase rate, reducing churn, strengthening the first-to-second purchase transition, or intervening with customers who are going quiet relative to their own pattern. Works on customers who are still reachable and still buying; use customer-winback once they have stopped.

Use when targeting customers who have already lapsed — dormant or churned contacts who have stopped buying and largely stopped engaging. Decides who is worth recovering, what would actually bring them back, how many attempts are justified, and when continued sending should stop in favour of suppression. Answers "should we keep emailing our dormant list?" and "can we win these customers back?". Use customer-retention for customers who are still active, and channel-optimization when the question is which channel to reach them on rather than whether to reach them at all.

Use when the question is whether the sending programme itself is fit to send — whether the sending domain authenticates and aligns, what the domain and IP reputation posture is, whether a new domain or a long silence means volume has to be ramped rather than resumed, and which mailbox providers are turning against this sender. Answers "are we authenticated?", "is our sender reputation getting worse?" and "can we safely increase volume?". Use list-hygiene when the remedy is which contacts to suppress, and email-quality-auditor when one specific campaign needs clearing before it goes out.

Use when deciding what a message should vary per recipient beyond their first name — which verified signal each dynamic element draws on, what it falls back to when that signal is missing, how deep the variation should go before it stops paying, and which elements should be dropped rather than defaulted. Answers "how should we personalize this email?", "what merge fields can we actually use?" and "what happens when the data isn't there?". Use content-optimization when the fixed wording is the problem, and product-recommendation-strategy when the question is only which items to show.

Use as the last check before a campaign is sent or an automation is switched on — sweeping audience, consent, suppression, recent contact, personalisation fallbacks, links, opt-out, offer accuracy, product availability, collisions with other sends and the sending programme's condition, then returning one verdict with every objection stated. Answers "can we safely send this?", "is this campaign ready?", "check this before it goes out" and "review this automation before we switch it on". Use campaign-optimization when a campaign has already run and underperformed, and deliverability-qa when the sending programme rather than one message is in question.

Use when a built email needs checking for whether every recipient can actually read it — how it behaves on a narrow phone screen, in dark mode, with images blocked, and for someone using a screen reader; whether alt text, contrast, link text and the plain-text part hold up; and whether the footer and opt-out survive. Answers "is this email ready to send?", "will this break in dark mode?" and "does it work with images off?". Use content-optimization when the wording itself is the problem, and email-quality-auditor for the whole pre-send gate of which rendering is one part.

Use when preparing a multi-message sequence around a holiday, festival or seasonal event — "prepare a Diwali campaign", "build our Black Friday sequence". Derives how many stages the sequence needs, which stages they are, how they are spaced, and which channel carries each, from the holiday's importance to this store, its sales window, product cycle, audience size and history. The stage count is derived, never fixed.

Use when deciding whether and how a store should participate in a holiday, festival or seasonal event — establishing relevance to this specific store, which products and audiences it applies to, what the strategy and timeline should be, and whether it is worth doing at all. Hand to holiday-drip-campaign to build the message sequence once the strategy is set.

Use when bounces, spam complaints or one mailbox provider turning against the store are the problem rather than the campaign — deciding which contacts leave the sending population and when, how to respond to a hard bounce versus a repeated soft one, whether a complaint trend means pausing sends entirely, what to do with subscribers who have never once opened, and when a sunset path beats another attempt. Answers "our bounce rate is climbing", "engagement has collapsed at one provider" and "who should we remove from the list?". Use deliverability-qa when authentication or sender reputation is the suspected cause, and customer-winback when lapsed buyers might still be recovered.

Use when the planning horizon is a year or a quarter rather than a single month — building the annual calendar, sequencing the year's major moments against each other, reserving capacity for known peaks, and keeping that long-horizon plan current as commitments change. Answers "plan our marketing for the year", "build a quarterly calendar" and "what does the next twelve months look like?". Use monthly-marketing-planner for a single period, which this skill composes.

Use when planning a month of marketing — "plan next month", building a marketing calendar, or deciding what campaigns to run over an upcoming period. Produces a dated calendar of campaigns with objectives, audiences, channels, offers, content direction, expected outcomes, dependencies and risks, derived from the store's own history rather than a fixed template. The number of campaigns is an output, never an input.

Use when deciding what incentive a message should carry, if any — whether a discount is needed at all, and if it is, whether a percentage, a fixed amount, free shipping, a spend threshold, a bundle, a gift, early access or exclusivity fits this audience and this objective better, and how deep it should go. Answers "what offer should we use?", "how much discount is right here?" and "do we need to discount at all?". Use aov-growth when the lever itself is still open between thresholds and merchandising, and campaign-optimization when a campaign has already run.

Use when nobody has named the problem — "what should we be working on?", "where should we focus?", "review our programme", or a sense that something is off with no specific symptom. Runs an open-ended scan across revenue, lifecycle leakage, automation coverage, campaign performance, list health and deliverability, channel mix, cadence, catalogue and content, then ranks whatever is most worth fixing or building next. Use revenue-growth when the objective is already framed as revenue.

Use when a new product, collection, restock or category is being introduced and needs a marketing plan — deciding who hears about it first, how the announcement is sequenced, which channels carry it, and how the launch converts into ongoing demand rather than a single spike.

Use when the question is which specific items to put in front of a given customer and on what evidence — choosing what fills a product block, ranking candidates for one contact rather than for the catalogue, deciding how far down the evidence ladder a recommendation may go before it is a guess, and saying how many items are worth showing. Answers "which products should we recommend?", "what should go in this email's product block?" and "how do we pick items per customer?". Use cross-sell when the objective is breadth into a new category, upsell when it is a move to a higher tier, and aov-growth when the lever itself is not yet chosen.

Use when products are consumed and rebought on a measurable cycle and the store should be reminding customers before they run out — deriving reorder intervals per product and size, deciding which products qualify, and timing the reminder ahead of run-out rather than after it. Use automation-architect for journey topology once the intervals are derived.

Use when someone asks what happened to revenue rather than what to do about it — decomposing revenue into its drivers, attributing it across campaigns, automations and channels, comparing periods, and explaining a rise or fall with evidence. Read-only. Use revenue-growth when the question is which opportunities to pursue.

Use when the objective is stated as revenue — "increase revenue this month", "we need to hit a number", "where is our growth going to come from". Identifies where the store's revenue can realistically grow, ranks the opportunities by evidence and expected value, and decides which campaigns, automations and audiences should carry them. Produces a strategy that other skills execute, not a calendar.

Use when the question is what hour and which day a message should go out — whether one send time for the whole list beats a per-contact time derived from when each person last engaged, whether this store has enough engagement history for per-contact timing to be anything other than noise, how local time zones and quiet hours bound the answer, and how to test a timing change without confounding it with a content change at the same time. Answers "when is the best time to send?", "does it matter what day we send on?" and "should we let it pick a time per person?". Use channel-optimization when the question is which channel carries the message, and automation-architect when the question is the delay between steps inside a journey rather than the hour a send goes out.

Use when a change in availability or price should reach specific customers rather than the whole list — a back-in-stock alert to the people who asked for one, a low-stock nudge on an item already sitting in a cart, a price-drop notice to people who viewed or saved something, or collecting a waitlist before a product exists. Decides who is notified and in what order, whether the units available justify notifying anyone at all, and when price-drop messaging starts teaching customers to wait for a discount instead of buying. Answers "set up back-in-stock alerts" and "should we announce that a price dropped?". Use product-launch when a product, collection or restock is being announced as a campaign, and product-replenishment for products bought again on a consumption cycle.

Use when onboarding a store onto TargetBay Email & SMS — newly signed up, just migrated or moved from another platform, or starting from scratch with no automations and no sending history. Answers "what should we set up first?" for email and SMS. Audits what data actually exists, decides the baseline journeys to build and in what order, plans a deliberately light first period, and sets a review point. Deliberately conservative, because most planning skills depend on history a new store does not have — this skill says so rather than inventing defaults.

Use when the mechanism is already chosen and it is a move upward — promoting a premium tier, a larger size, a bundle or a subscription to customers whose price-band behaviour shows headroom. Answers "promote our premium tier" and "who can we move up?". Use cross-sell when the goal is a purchase in a different category, and aov-growth when the lever is not yet chosen — it owns thresholds, merchandising and ranking the levers against each other.