The Autonomous Commerce is here

@Mercer build a Dark Roast Espresso Coffee Beans launch that feels like our brand. Give me a custom storefront with proper product photography, an editorial coffee story and a product page people can buy from. Use the new AETHEL campaign, warm ivory and espresso tones, and clear, confident typography.

Build a real DTC offer ladder: one bag at $32.95, two at $56.02 with 15% savings, and three at $74.14 with 25% savings. Show compare-at totals, per-bag value and the gifts each pack unlocks. Make the selected offer obvious and keep the purchase easy on mobile.

We have product-page, advertorial and listicle traffic. Check their actual destinations. A beautiful new PDP does not help if an old article still sends buyers to the retired page.

I’m building the customer experience around the product: a photographic gallery, clear bundle choices, a progressive gift unlock, useful tasting and brewing details, and a repeated purchase CTA. The design follows your brand brief; the page structure is custom.

The working page keeps espresso one-time and carries the selected pack straight into checkout. Catalog mapping and entry-route review stay in their own views.

The mapping preview will show paid units separately from the scoop and canister, so “two bags” cannot accidentally mean one paid bag plus a gift.

Exactly. First orders get a scoop with every pack; two or three bags also unlock the storage canister. The three-bag pack adds free US standard shipping. Show the unlocks before the CTA. A returning subscriber can buy espresso separately, but that must not change their subscription or give them another welcome gift. Show the two-bag and returning-customer cases before we approve it.

Merchandising now lets you inspect each pack against first-order and returning-customer eligibility. Two bags remain two paid units at $56.02, saving $9.88 against $65.90. The first order adds a $0 scoop and canister; the returning case adds no welcome gifts. Three bags total $74.14 with $24.71 savings and free US standard shipping. All packs stay one-time, with existing subscription terms untouched.

A multi-bag entry also skips the later two-bag stock-up offer.

Now check the listicle and the alternate landing page. One has an old destination in a custom content block. The other variant is still a draft. Do not enable a 50/50 split just because the main page is ready, and do not rewrite the live homepage while fixing those links.

Launch review separates those failures. The outdated listicle stops at the retired destination; its CTA needs the current route, followed by another quantity check. The unpublished variant stays out of traffic until its destination and checkout mapping are ready.

The working product CTA and advertorial path are separate cases. One passing path does not make the others ready.

Keep the useful detail in the review, but the page itself should still feel like a premium coffee brand. Make sure the campaign photos, visible price and selected pack survive on mobile. Check the real footer too—our custom sections have hidden menu links before.

The customer page stays focused on the product and purchase. From Merchandising, a compatible first-order pack opens that same selection on the page, then carries it into checkout. The returning-customer case remains separate so it cannot inherit the preview’s first-order gift.

The launch checklist retains rendered policy and Manage subscription links, every quantity and both variant destinations.

Draft ready ✅ The page, paid/gift mapping and launch routes can be reviewed independently. The original espresso order stays one-time; gifts never recur; multi-bag orders skip the stock-up. Stale links and unpublished targets remain visible decisions before traffic moves.

@Mercer carry the exact espresso offer into checkout: $32.95, $56.02 or $74.14 for one, two or three bags. Keep the first-order scoop and eligible canister as separate $0 lines. Three bags unlock free US standard shipping. No recurring charge. Keep tax separate until the full total is known. Preserve our processors and show wallets only when the device, domain and provider qualify.

I also need the failure paths. If the bank asks for authentication, an issuer declines or the shopper clicks twice, show me exactly what happens to the payment and order.

I’m separating the customer surface from the payment operation. The sample setup uses Stripe for the initial card purchase, exclusive Airwallex card rebills and PayPal for its own agreements. Smart 3DS belongs to the initial payment here.

The event inspector will distinguish customer authentication, payment success and a paid order instead of treating them as one status.

Good. If Airwallex is unavailable, do not silently push the rebill into Stripe just because the account is connected. And a provider token is not automatically portable. Show the unavailable-rail and incompatible-credential cases beside the healthy card renewal and PayPal agreement.

Payment rails now resolves those five operations. Healthy card renewals assume compatible saved credentials, recurring consent and the original reference. PayPal renewals stay on their agreement. The unavailable exclusive rail holds billing; the incompatible credential case stops before creating an attempt.

The provider and wallet details remain inspectable underneath.

For checkout, walk through the actual sequence. Passing 3DS cannot mean “paid” until the payment succeeds. A soft decline can offer a customer retry, but a hard decline should stop. Two clicks should not turn into two charges. And when you report conversion, one order with two attempts is still one converted checkout.

The staged event preview keeps those distinctions visible. After authentication completes, the order still waits for the payment result. A successful soft-decline retry shows two attempts and one paid order. The hard-decline path stays unpaid and asks for another method. A repeat submission keeps the original operation pending, then resolves one order.

Show post-purchase acceptance and decline too. The espresso order must stay paid either way. Accepting the Discovery subscription needs its own $24.95 every-30-days terms; a click alone is not another successful charge. If they decline or skip, there should be no extra payment.

The offer case starts with the original order paid. It then shows explicit recurring terms, acceptance with the offer payment pending, and only afterward the sample successful $24.95 product charge. That creates its own recurring agreement; espresso remains one-time.

Declining adds no charge and exposes the one-time rescue or skip path. The original order remains intact throughout.

Ready to inspect ✅ Checkout shows the customer, Payment rails explains the allowed processor and credential, and Paths & recovery follows seven event scenarios. Purchase-attempt counts remain separate from paid orders and offer charges. The store-wide eight-attempt recovery policy stays a separate draft; a scheduled retry is never counted as recovered revenue.

@Mercer check the purchase terms across this funnel, not just the policy page. The espresso order is one-time. Discovery is an optional subscription after checkout, and its decline branch is a one-time rescue. A shopper should never need to interpret the receipt to find out which one they accepted.

Show the item, amount and future charge at each touchpoint. Also keep the free welcome scoop out of paid quantities and future renewals.

I’ve separated six review contexts: the original espresso order, Discovery subscription, one-time rescue, first-order gift, delivery promise and returns. Each has the customer-facing wording, the connected touchpoints and any input that is still missing.

The espresso order stays $32.95 once. Accepting or declining a later offer cannot silently change that original purchase type.

The Discovery CTA currently just says “Add Discovery to my order — $24.95.” That reads like a one-time add-on. Make the renewal obvious without filling the page with fine print. And if they decline, the $28.01 alternative must not inherit the subscription language. Show me exactly what changes.

The proposed Discovery CTA says “Start my Discovery subscription — $24.95 today, then $24.95 every 30 days.” One bag per delivery. The same cadence belongs on the receipt and in the customer portal.

The rescue says “Add one Discovery bag for $28.01, one-time. No subscription or recurring charge.” Those branches are mutually exclusive. The consistency map lets you inspect each path without mixing their billing terms.

The scoop is a welcome gift, not a coffee unit and not the paid ceramic cup. I’ve seen that become “free with every delivery” in campaign copy. Keep it first order only.

Also remove the two-day delivery promise. We haven’t confirmed processing time or every shipping region yet. Don’t turn our dispatch target into an arrival guarantee.

The gift review separates the $0 scoop and eligible canister from paid coffee and the $12 cup. Proposed wording: “One complimentary welcome scoop with your first order. It is not part of future renewals.” Checkout and receipt keep it as its own line.

For fulfillment, the two-day guarantee is removed from the proposed copy. The tracking sentence remains; processing time, carrier transit, regions and charges are still merchant inputs. Publication stays held until those facts are confirmed.

Returns need the same care. We haven’t decided whether opened coffee qualifies, who pays return shipping, or the final window. Don’t borrow another store’s thirty-day policy.

Let me compare the actual before and proposed wording, then see where each change needs to appear. I want to approve the connected path, not three pages that contradict one another.

Change review keeps your selected context when you move from the consistency map. You can compare the original line with the proposed wording and expand the affected touchpoints.

For returns, the draft asks customers to contact support with an order reference. The window, opened-product eligibility, damaged or incorrect orders and return-shipping responsibility remain unresolved. It is a draft for review, not a completed return policy.

Ready for your review ✅ One-time and recurring charges are separated, the rescue cannot create a renewal, and the gift does not repeat with the subscription. Fulfillment and returns still show the facts you need to supply.

Nothing is published. Approve the actual wording and the matching purchase, receipt and support touchpoints before replacing the live copy.

@Mercer act as our CRO on the espresso funnel. The entry order is $32.95 for one bag, one-time. Build a three-step post-purchase flow that increases collected revenue and tests a complementary subscription. Do not offer months of the same coffee to someone who already bought two bags.

Check existing Discovery subscribers and customers who already have the tasting cup. I want the right offer for the actual order, with proper decline paths—not the same three cards for everyone.

I’ve drafted Discovery at $24.95 today and every 30 days. If recurring delivery is the objection, the alternative is the same bag at $28.01 once. Those are mutually exclusive. Then two espresso bags at $56.02, followed by the $12 tasting cup.

The subscription introduces a different roast; the stock-up extends the espresso supply; the cup adds a useful accessory. Each earns its place. The original paid order stays intact if the customer declines everything.

For an existing Discovery subscriber, skip the new subscription and its rescue—do not sell the same plan twice. A two-bag buyer skips the stock-up, and a cup already paid for must not appear again. Keep the free welcome scoop distinct from the paid cup.

On the rescue, say one-time clearly. A one-time alternative is not a cheaper recurring price.

Mapped all four entry paths. A one-bag buyer can see all three offers. A two-bag buyer skips stock-up. An existing Discovery subscriber skips both new enrollment and the one-time rescue. If the $12 cup is already paid for, it stays in the original $44.95 order and never appears again.

The rescue is a separate one-time purchase. Nothing converts the original espresso order into a subscription, and accepting the rescue cannot also create a recurring Discovery agreement.

Now test subscription-first against one-time Discovery. The traffic, eligibility, two-bag offer and cup stay the same. If B changes the stock-up discount as well, we will not know what caused the difference.

Use revenue per assigned eligible paid order, including people who never take an offer. Do not use only the people who saw or clicked the offer as B’s denominator.

The draft comparison changes only the opening offer. A begins with the $24.95 subscription and its decline rescue; B begins with the $28.01 one-time bag. Both keep the $56.02 stock-up and $12 cup, with the same exclusions.

Assign eligible paid orders once, 50/50. I’ve also included checks for mismatched subscriber audiences, a second discount change and an exposed-shoppers-only denominator. Any of those would spoil this comparison. Both flows need review and publication before the test starts—there is no winner yet.

Show me the money trail too. A shopper can accept while the saved-card charge is still pending or fails. That cannot count as upsell revenue or a successful subscription start. And someone can accept two offers—do not add those line counts together and call them unique converted orders.

Keep subscription starts separate from today’s cash. Future renewal value is not money we collected at checkout.

Here’s the reconciliation: 218 paid offer lines across 178 unique orders, $6,133.94 collected, and $7.67 per eligible paid order across 800 orders. A shopper accepting two offers is one converted order, with two paid lines. Each offer’s revenue contribution uses the same 800-order denominator; its paid take rate uses that offer’s views.

The separate payment example shows why acceptance is not enough. Pending or failed means $0 collected and no successful new subscription. Only the successful $24.95 charge counts. The sample’s 68 Discovery starts have $24.95 every-30-days terms; future renewals are not included in collected revenue.

Ready for review ✅ Every offer shows its price and purchase type before acceptance. Declining Discovery opens the one-time rescue; declining that still reaches the next eligible offer. Skip remaining offers preserves earlier choices and the original paid order.

Before activation, review the four entry contexts, walk the one- and two-bag customer paths, and approve both experiment drafts. The commercial rules are ready to inspect; no live traffic, original order or payment has been changed.

@Mercer rebuild recovery around the actual shopper. One person left a $32.95 espresso bag; another left the discounted two-bag bundle at $56.02; someone else chose a subscription. They should not all get a coupon. Keep helpful, discount-free reminders at 4, 24 and 72 hours from checkout start.

Recheck purchase, consent and replacement checkouts before every slot. Our other tool owns subscription receipts—leave those alone. I also want a separate replenishment campaign for one-time buyers, not another abandoned-cart message.

Recovery and replenishment need different audiences and timing. Recovery follows an unpaid checkout; the campaign starts with a reviewed buyer audience. I’m keeping those drafts separate.

For recovery, purchase or withdrawal of consent stops the message before an incentive is evaluated. Subscription carts retain the helpful reminder without a coupon. The saved pack and purchase type stay attached to the checkout.

At 72 hours only, propose a 50/50 test against the current no-discount copy: 10% off, $5 cap, $30 minimum. First order, one-use code, no stacking, 24-hour expiry from send. Keep it unpublished.

The two-bag bundle already has 15% off, so exclude it from another incentive. Show that alongside the difficult recipients: bought already, unsubscribed, returning customer, subscription cart, or received an incentive seven days ago.

The one-bag challenger saves $3.30, leaving a $29.65 item subtotal. The two-bag bundle stays at $56.02: its existing 15% savings block an additional code. Shipping and tax stay separate. The 4- and 24-hour messages and the 72-hour control remain incentive-free.

The seven-case recipient review shows who exits, who keeps the standard reminder and who qualifies for the proposed challenger. Eligible one-time examples open the matching email and checkout preview.

Now follow the timing. They buy at hour six after the first reminder—remove everything pending. Or they start a new two-bag checkout at hour ten. Don’t send reminders for both carts or hit them twice inside twelve hours.

For replenishment, exclude active subscribers and people who purchased in the last thirteen days. Draft email and SMS, but an email opt-in must never count as text-message permission.

The recovery timeline shows the purchase exit and the replacement-cart branch. In the replacement example, the new 4-hour slot lands at hour 14; the 12-hour frequency limit moves it to hour 16. Later slots re-anchor to hours 34 and 82, with eligibility checked again.

The campaign draft has four recipient reviews and separate email/SMS copy. The requested rolling 13-day exclusion is not supported by the current all-time order filter. I’ve kept it as an audience-review requirement; any external audience or import workflow needs verification before scheduling.

Good. Keep the campaign as a one-time reorder: the right espresso pack, no surprise subscription, no fake deadline. Recent buyers and active subscribers should not receive this draft; contacts without SMS consent stay out of the text version.

Before release, show the sender problems separately from the content. I don’t want a polished draft to imply the domain or footer is already approved.

Readiness separates DNS, sender and reply handling, footer details, the saved checkout and message ownership. This draft still has a shortened DKIM value, three missing records and an unconfirmed business address. Existing subscription receipts remain with their current sender.

The creative preview keeps the selected pack, message timing and test arm together. Campaign exclusions are review decisions only; no live audience has changed. Sending stays held.

Ready to inspect ✅ Seven recovery recipient cases, four event timelines, a separate email/SMS campaign brief and the launch gates are connected. The campaign draft distinguishes older one-time buyers, recent buyers, subscribers and email-only permission.

No message is scheduled. DNS and footer approval, audience verification and the experiment decision are still required. A coupon redemption would be attributed activity—not proof that the offer created incremental revenue.

@Mercer separate the subscription problems before we send another payment email. We replaced the House Blend options last week, and some legacy plans stopped renewing. Separate a deleted mapping, an issuer decline and an unavailable rebill rail. Keep their $48 two-bag price and 30-day cadence. Do not collect the missed renewal while you investigate.

The diagnosis follows agreement → catalog → renewal → payment. A deleted mapping stops before any charge; a soft decline reaches the issuer; a hard decline exits retries. An inactive exclusive Airwallex rail holds billing without rerouting. The first case points to a deleted option. Was it recreated rather than edited?

Yes, deleted and recreated. There are two replacements with nearly the same name: a one-bag option and the new $52 two-bag option. Neither should rewrite the customer’s $48 agreement. Show me all three repair proposals so I can see what you would reject, not just the happy path.

The equivalent two-bag mapping passes. Mapping the one-bag option fails the quantity check; adopting $52 fails the legacy-price check. The accepted proposal changes only the catalog reference. Two bags, $48, every 30 days and the saved payment method stay intact. No overdue charge is queued.

That’s the distinction I wanted. For actual card failures, use eight attempts over 17 days: three at full price, then 10%, 15%, 20%, 25% and 30% off that charge only. Hard declines should leave the sequence and request a payment update. Is that policy limited to these House Blend plans?

The policy is store-wide, so it would apply to every selling plan. The shared draft schedules days 1, 2, 4, 6, 8, 11, 14 and 17. On this $48 plan, the final attempt is $33.60; future renewals stay $48. Hard declines skip retries and pause, with a secure payment-update email.

Keep mapping approval separate from the store-wide retry draft. Then show each payment outcome: recovered, soft decline again, hard decline, and all eight exhausted. If day 17 succeeds at $33.60, stop the remaining recovery messages and keep the next regular renewal at $48. Scheduled retries must not show up as collected revenue.

Select an attempt, then its outcome. Success records only that charge and stops this renewal’s retry sequence. A soft decline moves to the next proposed day; a hard decline stops immediately. A failed eighth attempt pauses the plan and sends the paused email. The agreement stays $48, and the missed renewal remains a separate billing decision.

Ready for review 👍 Diagnosis distinguishes no payment attempt from an actual failed charge. Repair shows exactly why the one-bag and $52 proposals are blocked. Recovery connects the amount to the queue, customer message and collected revenue. The mapping proposal is prepared; the policy remains a draft, with no catch-up charge scheduled.

@Mercer rebuild our cancellation journey around why subscribers leave. Too much coffee should get a slower cadence first, then a smaller pack—not a coupon. Price objections can get 15% off one renewal, once per subscriber. Keep discounts from stacking. I want accepted offers versus customers who actually paid their next renewal.

Oversupply gets a cadence change first: two bags every 45 days, still $48. The price path needs its own eligibility checks. I suggest 2+ paid renewals, a 90-day incentive cooldown, no other discount queued and no previous use of this 15% save. That prevents the same subscriber from collecting it repeatedly.

Yes. Test a customer with one paid renewal, someone who got a save offer last month, and someone with their 10% loyalty reward already queued. Don’t take that earned reward away just to show a retention coupon. And a pause or failed card should never reset—or advance—their reward progress.

The decision preview now explains each result. One renewal fails the minimum; last month’s offer fails the cooldown; a queued loyalty reward blocks stacking. That customer keeps the earned 10%, so the next product charge stays $43.20. The eligible case gets $40.80 once, then returns to $48.

Someone who used the 15% save four months ago still fails the once-per-subscriber rule. Keep non-discount choices and quote the actual smaller-pack price. Now reconcile the reporting: how do 108 accepted saves become only 72 paid renewals? Show pending and later-lost customers, and don’t count a support request or an abandoned journey as a save.

Prior use stays blocked; cadence, a quoted pack swap or a one-cycle skip remain available. Of 240 attempts, 108 accepted a save: 72 paid, 28 are not due yet and 8 were later lost. The other 12 reached support, 30 made no choice and 90 cancelled. Inspect each outcome’s event history. The $3,456 is paid revenue after retention, not a claim of incremental lift.

Now connect that to the rewards system. Our customer has four paid renewals. Show a failed payment, a pause, then two successful renewals. The free bag after six should be its own fulfillment order. I also want 10% after renewal three and the weighted mystery reward after nine, with milestone performance beside the customer ledger.

The ledger stays at four on a failed payment or pause, moves to five after one successful renewal, and reaches six after two. At six, the free House Blend bag becomes a separate $0 fulfillment order with a reward email or connected Klaviyo event. The regular renewal remains $48. The dashboard separates milestone progress from completed rewards.

Ready to inspect 👍 Flow explains who qualifies and why; Rewards follows each paid event into progress and fulfillment. Overview traces all six outcomes without erasing later losses from the original save count. Customer preview keeps the next charge, future terms and cancellation choice visible before confirmation. Everything remains in draft.