VIP and Loyalty Segment Campaigns

Most retailers have a loyalty program and no loyalty strategy. There is a points balance somewhere, a birthday email, and a monthly newsletter that goes to everyone on the list at the same time with the same offer. The customer who spent heavily across eleven orders last year gets the identical twenty-percent-off message as the person who bought one discounted item eighteen months ago and has not opened an email since. Both messages are wrong. The first discounts a customer who would have paid full price. The second wastes a send on someone who is gone.

The money leaks in two directions at once. On the top end, margin is given away to people who did not need the incentive, because the campaign has no way to know who they are at send time. On the bottom end, the genuinely lapsed are never worked hard enough, because they sit in the same list as everyone else and receive the same gentle newsletter rather than a real win-back. In between, the most valuable group in any retail database - the customers who bought twice and are deciding whether to buy a third time - get no special attention at all, even though that second-to-third purchase is where a one-off buyer becomes a habit.

The reason it stays broken is that proper segmentation looks like a data project. Someone has to define value tiers, calculate them for every contact, keep them current as behavior changes, and then build different campaigns per tier - and that is more work than any marketing calendar has room for. This build makes the segmentation a standing piece of infrastructure rather than a project. Scores are recalculated automatically, tier membership moves on its own, and each tier has its own campaign cadence, offer policy and escalation path. You stop building lists and start maintaining rules.

Pipeline and stages

Two pipelines do different jobs here. The first tracks the lifecycle of the relationship, the second tracks individual high-value opportunities that deserve human attention.

Customer Lifecycle - one opportunity per contact, moved by score, not by hand:

• Prospect - No Purchase - on the list, never bought.
• First Purchase - one order, inside the activity window.
• Repeat - Building - two or three orders. The most important stage in the pipeline.
• Core - regular buyer, four or more orders, recency healthy.
• VIP - top value decile by trailing twelve-month spend.
• At Risk - recency has slipped past the category's normal repurchase gap.
• Lapsed - well past the window, win-back territory.
• Dormant - Suppressed - no engagement after full win-back. Removed from regular sending to protect deliverability.

VIP Concierge - only for contacts in the VIP stage, worked by a named person:

• New VIP - Welcome Due, Active - Nurtured, Preview Invited, Personal Outreach Open, VIP At Risk, Retained, Churned VIP.

Calendars

• VIP Personal Shopping - bookable 45-minute appointments offered only to the VIP segment, in-store or video.
• Preview Event Slots - limited capacity sessions for new-range previews, with VIP-only booking links.
• Concierge Call Back - 15-minute slots on the account manager's calendar, used when a VIP goes quiet.

Custom fields

These are the engine. Everything downstream reads them, so get the names and the refresh logic right before building campaigns.

• Lifetime Order Count - all-time paid orders.
• Trailing 12m Spend - rolling twelve-month total, recalculated monthly. The primary value measure, because all-time spend never lets anyone fall.
• Average Order Value - trailing twelve months.
• Last Purchase Date and Days Since Last Purchase - recency, the single strongest churn predictor in retail.
• Purchase Interval Average - mean days between orders for this contact. At Risk is defined against this, not against a flat company-wide number, so a customer who buys every six weeks is not chased on the same clock as one who buys twice a year.
• Category Affinity Primary and Category Affinity Secondary - most-bought departments, used to pick which range to feature.
• Discount Dependency - share of orders placed with a code. High dependency means never lead with full price; low dependency means never lead with a discount.
• Return Rate - returns as a share of items bought. A high-spend, high-return customer is not automatically a VIP.
• Engagement Score - opens, clicks, SMS replies and site visits over 90 days, recalculated weekly.
• Value Tier - the calculated output: prospect, first, building, core, vip.
• Tier Changed Date - when membership last moved, so you can suppress whiplash messaging.
• VIP Entry Date and VIP Review Date - VIP status is reviewed, not permanent.
• Preferred Channel - email, SMS or WhatsApp, inferred from response behavior and overridable by the customer.
• Concierge Owner - the named staff member for VIPs.
• Offer Eligibility - full price only, standard promotion, or deep discount. The guardrail that stops a VIP being sent a clearance code.

Tags

• tier-prospect, tier-first, tier-building, tier-core, tier-vip - mutually exclusive, maintained by the scoring workflow. Never applied by hand.
• risk-at-risk, risk-lapsed, risk-dormant - the recency axis, which moves independently of value.
• vip-concierge-assigned, vip-preview-invited, vip-personal-shopper-booked.
• behave-full-price-buyer, behave-discount-led - drives offer selection.
• behave-high-returns - excludes the contact from VIP promotion regardless of spend.
• affinity-{department} - one per department, applied from Category Affinity Primary.
• suppress-promotional - a hard suppression flag that every campaign checks first.
• winback-attempted with a cycle number, so a contact is never put through the same win-back twice.

Forms and surveys

• Preference Center - channel, frequency, and department interests. Writes to Preferred Channel and the affinity tags. Linked in the footer of every send, because a customer who self-selects out of three topics is worth more than one who unsubscribes from everything.
• VIP Welcome Profile - sizes, favorite brands, occasions, and a free-text note the concierge can actually use. Sent once on entry to VIP.
• Win-back Reason - one question, four options: too expensive, did not need anything, bad experience, shop elsewhere now. The single most useful piece of data in the whole build, and almost nobody collects it.
• Post-purchase Pulse - two questions after the second order, feeding Engagement Score and flagging dissatisfaction before it becomes churn.

BO's role and guardrails

• In scope: answering "what tier am I in" and "how do I get the VIP benefits" from the contact's own fields, taking preference changes and writing them to the Preference Center fields, booking VIP personal shopping and preview slots from the calendars, and capturing win-back reasons conversationally when a lapsed contact replies to a message.
• Out of scope and escalated: any request for a discount not already in the contact's Offer Eligibility, any complaint, any query about a specific order or return, and any VIP inbound where the Concierge Owner is set - those route to the named human every time.
• Guardrails: BO never invents or negotiates an offer, never states a points balance or spend figure unless the field is populated and current, never tells a customer their tier is about to change, and never promotes a contact between tiers. Tier movement belongs to the scoring workflow alone.

Numbers and channels

• Email carries the segmented campaigns - it is the only channel with the room for a properly merchandised message.
• SMS is reserved for VIP and Core: early access windows, concierge replies, and preview invitations. Sending promotional SMS to the lapsed tail is how a number gets a bad reputation.
• A dedicated concierge number or shared inbox for VIP, answered by a human during trading hours, with BO covering out of hours and handing over with a summary.
• Segment audiences synced outbound to the ad platforms for suppression and lookalike building, which is where reporting-attribution earns its place in this build.
• Internal Slack channel for VIP At Risk alerts, because that notification needs to be seen within hours, not at the next weekly review.

1. A customer places her first order. The order webhook fires, the contact is created or matched, and the nightly scoring workflow sets Lifetime Order Count to 1, stamps Last Purchase Date, writes Category Affinity Primary from the line items, and sets Value Tier to first. The tag tier-first is applied and an opportunity is created in Customer Lifecycle at First Purchase.
2. Because she used no discount code, Discount Dependency is 0 and behave-full-price-buyer is applied. Her Offer Eligibility is set to full price only. She will never be sent a blanket clearance code while that holds - the first margin decision the build makes on its own.
3. Fourteen days later the First-to-Second nurture runs. She gets a two-email sequence built around her affinity department, with no discount: care instructions for what she bought, then a styling or use-case email featuring the natural companion product.
4. She buys again at day 31. Lifetime Order Count becomes 2, Purchase Interval Average is set to 31 days, and the tier recalculation moves her to building. The opportunity moves to Repeat - Building and the Second-to-Third sequence starts, which is the most important campaign in the build.
5. The Post-purchase Pulse goes out after order two. She rates the experience well and selects two department interests, which apply affinity- tags and raise Engagement Score.
6. Over the next seven months she places five more orders. On the monthly recalculation her Trailing 12m Spend crosses into the top value decile, behave-high-returns is absent, and Return Rate is within tolerance. The tier workflow promotes her to vip, swaps tier-core for tier-vip, stamps VIP Entry Date, sets VIP Review Date twelve months out, assigns a Concierge Owner by round robin, and creates an opportunity in VIP Concierge at New VIP - Welcome Due.
7. The VIP welcome fires: a personally signed email from the named concierge - not from the brand - introducing herself, the VIP Welcome Profile form, and the standing benefits. No discount is mentioned, because this customer has never needed one. A task lands on the concierge to send a one-line personal message within 48 hours.
8. A new range launches. The Preview Invitation workflow sends the VIP segment a 72-hour early access window with a Preview Event Slots booking link before anything goes to Core or below. She books a slot; vip-personal-shopper-booked is applied and the appointment lands on the concierge's calendar.
9. Four months later she stops buying. When Days Since Last Purchase exceeds twice her own Purchase Interval Average, the risk workflow applies risk-at-risk and moves the Concierge opportunity to VIP At Risk. Because she is VIP, no automated win-back discount is sent. Instead a Slack alert and a same-day task go to the concierge: call, do not email.
10. The concierge calls, learns she has moved house and changed her routine, and books a video personal shopping session. The call outcome is logged, the opportunity moves to Personal Outreach Open, and the automated at-risk sequence is suppressed for 30 days so the human conversation is not stepped on by a template.
11. She buys again. Last Purchase Date updates, risk-at-risk is removed, the Concierge opportunity moves to Retained, and a retention note is written against her record with the reason - the data point that makes next quarter's retention review useful rather than anecdotal.
12. Elsewhere in the database, a contact who bought once eighteen months ago hits the lapsed threshold. He gets the three-step win-back with escalating incentive and the Win-back Reason single-question form. He answers "shop elsewhere now". The answer is stored, winback-attempted cycle 1 is applied, and he is excluded from win-back for twelve months rather than being cycled again next quarter.
13. He never engages. After the win-back closes with no open and no click, the suppression workflow applies risk-dormant and suppress-promotional and moves him to Dormant - Suppressed. He stops being sent to, which lifts list-wide open rates and protects the sending domain - a loyalty program's least glamorous and most valuable function.
14. On the first of each month, the owner opens the segment dashboard: contact counts and trailing spend by tier, movement between tiers over the period, and campaign revenue attributed by segment. The board shows which tier is growing, which is leaking, and whether the Second-to-Third sequence is still earning its place.

1. Nightly score refresh

Trigger: scheduled daily at 02:00 local, plus an immediate run for any single contact on order webhook so a fresh purchase is never stale by a day.

Conditions: contact has at least one order, or is being initialized as a prospect.

• Recalculate Lifetime Order Count, Trailing 12m Spend, Average Order Value, Last Purchase Date, Days Since Last Purchase, Purchase Interval Average, Return Rate and Discount Dependency.
• Recalculate Category Affinity Primary and Secondary from the trailing twelve months of line items, weighted by spend not unit count, and sync the matching affinity- tags.
• Recalculate Engagement Score from 90-day opens, clicks, SMS replies and site sessions.
• Set behave-full-price-buyer when Discount Dependency is under 20 percent, behave-discount-led when over 60 percent, neither in between. Set behave-high-returns above your category tolerance.
• Derive Offer Eligibility: full price only for full-price buyers in the top two tiers, standard promotion for most, deep discount only for lapsed and dormant contacts.

Stop conditions: skip contacts flagged as test records or staff accounts; log and skip any contact whose order feed produced an impossible value rather than writing a corrupt score.

2. Tier assignment, promotion and demotion

Trigger: completion of workflow 1 for a contact, plus a full monthly pass on the first of the month to recompute decile boundaries across the whole base.

Branches: evaluated in order, first match wins.

• No orders ever, on the list - tier-prospect, stage Prospect - No Purchase.
• Lifetime Order Count equals 1 - tier-first, stage First Purchase.
• Count 2 or 3 - tier-building, stage Repeat - Building.
• Count 4 or more and Days Since Last Purchase under twice Purchase Interval Average - tier-core, stage Core.
• Trailing 12m Spend in the top decile, Return Rate within tolerance, behave-high-returns absent - tier-vip, stage VIP.

• On any change: remove the old tier tag, apply the new one, stamp Tier Changed Date, and write a note recording the old and new tier.
• Promotion guard: do not promote a contact whose Tier Changed Date is within the last 14 days. Tier whiplash makes the messaging look broken.
• Demotion guard: VIP demotion only runs at VIP Review Date, never mid-year, and triggers a review task for the Concierge Owner rather than a silent downgrade. No demotion message is ever sent to the customer.
• On promotion to vip: stamp VIP Entry Date, set VIP Review Date twelve months out, assign Concierge Owner by round robin, apply vip-concierge-assigned, create the VIP Concierge opportunity at New VIP - Welcome Due.

Stop conditions: exit without change if the computed tier matches the current tier; never move a contact carrying suppress-promotional into an active tier without a human releasing the suppression.

3. First-to-second purchase nurture

Trigger: tag tier-first applied.

Conditions: marketing consent present, suppress-promotional absent. Branch on Category Affinity Primary to choose the content block.

• Wait 3 days, then send email 1: "Getting the most from your {{product category}}". Pure value - care, setup or use guidance. No offer. The purpose is to earn the second open.
• Wait 11 days, then send email 2 featuring the companion product for their affinity department: "What people pair with {{Category Affinity Primary}}". Still no discount if behave-full-price-buyer is present.
• Discount-led branch: if behave-discount-led is present, email 2 carries the standard promotion instead, because this customer's demonstrated behavior is to wait for one.
• Wait 14 days. If no second order, send email 3: a single social-proof-free reminder of the range, plus the Preference Center link so a disengaged contact can narrow rather than leave.
• Wait 30 days. If still no second order and Engagement Score is zero, stop sequencing this contact and let the recency workflow pick them up.

Stop conditions: second order placed, which immediately cancels remaining steps and hands off to workflow 4; unsubscribe; tier change.

4. Second-to-third purchase program

Trigger: tag tier-building applied.

Conditions: marketing consent present. This is the highest-value sequence in the build and should be reviewed monthly rather than left to run for a year untouched.

• Immediately send the Post-purchase Pulse - two questions, no offer. A low score branches to a service recovery task for a human within 24 hours rather than continuing the sequence.
• Wait 7 days, then send email 1: "You are two for two" - acknowledge the repeat explicitly, introduce the loyalty mechanics and what the next tier gives them. Naming the ladder is what makes people climb it.
• Wait 14 days, then send email 2 built on Category Affinity Secondary, deliberately widening the range they consider rather than selling the same department again.
• Wait 21 days. If no third order, send SMS - only if Preferred Channel is SMS and consent allows: "{{Business name}}: your {{affinity}} picks are back in. Have a look: {{link}}"
• Wait until Purchase Interval Average plus 7 days from Last Purchase Date, then send email 3 timed to their own rhythm rather than a calendar. This timing step is the difference between a sequence and a campaign.

Stop conditions: third order placed; tier moves to core or vip; dissatisfaction flagged by the Pulse, which hands the contact to a human and suppresses promotional steps for 14 days.

5. VIP welcome and concierge onboarding

Trigger: VIP Concierge opportunity created at New VIP - Welcome Due.

Conditions: Concierge Owner is populated. If not, hold and alert the manager - an unowned VIP is worse than no VIP program.

• Send email immediately, from the concierge's own name and reply address: "Hello {{first name}}, I am {{Concierge Owner}} and I look after our best customers. You now have early access to new ranges, a direct line to me, and personal shopping whenever you want it. If you tell me a little about what you like, I can be useful rather than generic: {{VIP Welcome Profile link}}" No discount, no points language.
• Create a task for the concierge due in 48 hours: "Send a personal one-liner to {{first name}} - reference {{Category Affinity Primary}}." The automation sets up the relationship; a human has to start it.
• On VIP Welcome Profile submission, write the preferences, notify the concierge, and move the opportunity to Active - Nurtured.
• Wait 7 days. If the form is not submitted, send one SMS from the concierge number: "{{Concierge Owner}} here from {{Business name}} - anything you are looking for at the moment?" Then stop asking.
• Add the contact to the VIP audience sync for ad suppression, so you are not paying to acquire someone you already have.

Stop conditions: contact leaves the vip tier; unsubscribe; concierge marks the relationship as do-not-contact.

6. VIP and Core early access

Trigger: manual launch of a new range or event, selecting the campaign.

Conditions and sequencing: the whole point is the order of the waves.

• Wave 1 - VIP, hour 0: email plus SMS where Preferred Channel is SMS. "{{first name}}, the new {{range}} is live for you 72 hours before anyone else. Book a preview slot with {{Concierge Owner}}: {{Preview Event Slots link}}" Apply vip-preview-invited.
• Wave 2 - Core, hour 72: email only. Early access framing, no concierge booking link.
• Wave 3 - building and first, hour 120: general launch email.
• Excluded entirely: any contact carrying suppress-promotional, risk-dormant, or lacking marketing consent. Lapsed contacts are excluded from launches and handled by workflow 8 instead - a launch email is a poor win-back.
• Wait 24 hours after each wave, then send one reminder to non-openers in that wave only, never to the whole segment.
• Create a task for the concierge listing VIPs who opened but did not buy, for personal follow-up within 72 hours.

Stop conditions: purchase removes a contact from all later waves and reminders; stock-out ends the campaign for every unsent wave.

7. At-risk detection and recovery

Trigger: nightly evaluation - Days Since Last Purchase exceeds twice Purchase Interval Average, with a floor of 45 days so a weekly buyer is not flagged after two weeks.

Branches on tier:

• VIP branch: apply risk-at-risk, move the Concierge opportunity to VIP At Risk, post to the Slack alert channel, create a same-day call task for the Concierge Owner, and send no automated message at all. Suppress all promotional sequencing for 30 days to give the human room.
• Core branch: apply risk-at-risk. Wait 2 days, then send email: "It has been a while, {{first name}}" featuring new arrivals in their affinity department, no discount. Wait 10 days; if no purchase and no open, send the standard promotion.
• Building and first branch: apply risk-at-risk and run a two-step sequence with the standard promotion in step two.
• All branches: if a purchase occurs, remove risk-at-risk, write a recovery note with the channel that worked, and move any Concierge opportunity to Retained.
• If no purchase after 60 days, apply risk-lapsed, move the lifecycle opportunity to Lapsed, and hand to workflow 8.

Stop conditions: purchase at any point; concierge logs a reason to hold, which pauses the workflow for a stated period; opt-out.

8. Win-back, reason capture and suppression

Trigger: tag risk-lapsed applied.

Conditions: no winback-attempted tag from the last twelve months. A contact goes through this once a year, not every quarter.

• Apply winback-attempted with the cycle number. Set Offer Eligibility to deep discount for the duration of this sequence only.
• Day 0, email 1 - no offer, one question: "Are we still your kind of shop? {{Win-back Reason link}}" The answer is worth more than the sale.
• Day 7, email 2 - the standard promotion, with the affinity department front and center.
• Day 21, email 3 - the final and strongest offer, with a stated expiry date and the honest line: "If we are not for you any more, that is genuinely fine - {{unsubscribe link}}. Otherwise, here is our best."
• Reason branches: "bad experience" routes immediately to a human with the full order history and stops the sequence. "Too expensive" tags the contact for outlet and clearance campaigns only. "Did not need anything" keeps them in low-frequency seasonal sends. "Shop elsewhere now" moves them straight to suppression without the remaining steps.
• Day 35, evaluate. Any purchase returns the contact to the normal lifecycle with Offer Eligibility reset. No open and no click across all three sends applies risk-dormant and suppress-promotional and moves them to Dormant - Suppressed.
• Quarterly, export the dormant population for ad-platform suppression and remove them from all regular sending.

Stop conditions: purchase; unsubscribe; reason answered as shop elsewhere; twelve-month cooling period blocks re-entry.

9. BO preference and tier inquiry handler

Trigger: inbound SMS, WhatsApp or chat message containing a loyalty, tier, benefits or preference intent.

Branches:

• Tier or benefits question: BO reads Value Tier and states the benefits for that tier verbatim from a fixed benefits block. BO does not estimate how close the customer is to the next tier, because that invites a negotiation it cannot have.
• Preference or frequency change: BO sends the Preference Center link, or takes the change conversationally and writes Preferred Channel and the affinity tags directly, confirming what it changed.
• Booking intent: if tier-vip is present, BO offers slots from VIP Personal Shopping or Preview Event Slots and books. If the tier is not vip, BO offers the standard booking route without mentioning the VIP calendars.
• Discount request: BO checks Offer Eligibility and shares only an offer already assigned to that contact. It never creates, extends or improves one. Anything beyond that goes to a human: "Let me ask the team and come back to you."
• Any VIP inbound where Concierge Owner is set: BO answers only out of hours, and in hours assigns straight to the concierge with a one-line summary.
• Complaint, return or order query: immediate escalation with a summary note, no attempt to resolve.

Stop conditions: BO stands down for the rest of the thread once a human replies; BO never sends more than two consecutive messages without a customer response.

Before this build, "the list" is one object and every campaign is a compromise between the people on it. After it, the list is five populations with different economics, and each campaign is addressed to one of them. Margin stops leaking at the top because full-price buyers are structurally excluded from discount sends by Offer Eligibility rather than by somebody remembering. Revenue appears at the bottom because the lapsed are worked with a real win-back and then suppressed, instead of being newslettered forever at the cost of your deliverability.

The operational change is that nobody builds a list again. Marketing picks a segment that already exists and already contains the right people, which collapses the lead time on a campaign from days to an hour. The strategic change is that the second-to-third purchase gets its own program, which is where most retail databases are quietly bleeding.

Measure these monthly, and insist on them by tier rather than in aggregate:

• Tier distribution and movement - counts per tier and net promotions versus demotions. Healthy looks like a widening Core and a VIP population that grows slowly and loses slowly.
• Repeat purchase rate at order two and order three - the two conversions this build exists to move.
• Trailing 12m spend per tier and spend concentration - what share of revenue the VIP tier carries. Rising concentration is a revenue win and a risk worth naming.
• Revenue per send by segment - the number that proves segmentation beats blasting. Compare against your pre-build baseline, not against a benchmark.
• Discount share of revenue, split by tier. The intended result is that discount-driven revenue falls in the top two tiers and holds or rises only where it is doing real work.
• At Risk recovery rate - contacts carrying risk-at-risk who purchase within 60 days, split between automated recovery and concierge-led recovery, so you know whether the human touch is worth its cost.
• Win-back reason distribution - your cheapest ongoing source of truth about why customers leave.
• Suppression volume and list health - contacts moved to Dormant, plus open rate and bounce rate before and after suppression.
• VIP retention at review date - share of VIPs still VIP twelve months after VIP Entry Date.

Data you must have before scoring can work

• A reliable order feed with order date, order total, line-level department or category, and discount code usage, flowing into the platform from the commerce system or EPOS. Shopify, WooCommerce and BigCommerce handle this natively; most EPOS systems need Zapier or Make in between. Without discount code data you cannot compute Discount Dependency, and the offer guardrails become guesswork.
• Returns data, or at least a returns flag per order. Promoting a high-return customer to VIP is an expensive mistake and the only defense is the field.
• At least twelve months of order history imported before launch. Tier thresholds calculated on three months of data will be wrong, and the build's credibility with the client dies the first time a genuine VIP is addressed as a prospect.
• Identity resolution decided up front: how you match the same human across web orders, in-store transactions and the email list. Duplicate contacts split spend and silently demote your best customers. This is the single most common reason segmentation builds fail.
• Agreed tier definitions in writing. Decide whether VIP is a top decile, a spend threshold, or a committee decision, and whether demotion is allowed. Retailers often discover they disagree internally at exactly this point, which is better than discovering it after launch.

Integrations

• Commerce platform or EPOS for orders and returns.
• Ad platform audience sync for suppression and lookalikes, where paid media is part of the mix.
• Google Calendar or similar two-way sync for the concierge and personal shopping calendars.
• Slack or an equivalent channel for VIP At Risk alerts.
• Accounting or reporting export if spend-by-tier needs to reconcile with the finance view.

Numbers and messaging compliance

• Provision SMS numbers and complete A2P 10DLC registration in the United States before sending any promotional SMS. Marketing traffic faces stricter carrier scrutiny than transactional, so register the campaign honestly as marketing and expect the review to take time. Confirm equivalent sender registration requirements in other markets.
• Promotional sending requires marketing consent, and consent is not implied by a purchase in most jurisdictions. Keep the consent state on the contact, check it as the first condition in every campaign workflow, and never let a lifecycle workflow send promotional content to a contact who only consented to transactional messages.
• Opt-out handling on every channel: STOP, UNSUBSCRIBE and STOPALL on SMS, a working one-click unsubscribe on email, and a Preference Center link in every footer so the choice is granular rather than all-or-nothing. Honor the opt-out globally and immediately - a suppressed contact must not be reachable by any tier campaign.
• Quiet hours enforced per workflow and per time zone. Early access windows and preview invitations are the messages most likely to be scheduled at an inconvenient hour in pursuit of urgency. Hold them to trading hours.
• Keep the VIP channel genuinely low volume. Over-sending to the top tier is the fastest way to turn your best customers into unsubscribes.

Data protection

• This build creates profiling: scores, tiers and behavioral inferences about identifiable people, used to decide what they are offered. Several regimes treat that as a category requiring a stated lawful basis, transparency about the logic, and in some cases a right to object. Document what you calculate and why before you switch it on.
• Keep the VIP Welcome Profile free of anything you do not need. Sizes and brand preferences are commercial; health, family or financial detail volunteered in a free-text note is not yours to retain.
• Set retention rules for behavioral data and a deletion path that removes scores and tags, not just the contact record.
• Be able to answer, in one sentence, why a given customer received a given offer. If you cannot, the segmentation is too opaque to defend.
• This is operational guidance, not legal advice. Confirm your lawful basis, profiling disclosures, consent wording and retention periods with the client's own counsel or data protection adviser before launch.

People and sequencing

• A named concierge, or two, with real capacity. VIP Concierge is a human pipeline; building it without staffing it produces an abandoned board and a worse experience than no program at all.
• Build in this order: data in, scoring and tiers correct and verified against a hand-checked sample of twenty contacts, then suppression, then campaigns. Teams that build campaigns first spend the next month debugging audiences.
• Tier thresholds, the identity-resolution approach and the offer policy are the parts that differ most by retailer. Book a call at bookings.appbo.io to scope yours.