E-commerce
July 1, 2026
"Why are you offering me baby products when I don't have any children?" "You're recommending this serum again that I bought two months ago." "Your 'you will also like' emails have nothing to do with my order." Three messages where a shop with recommendation widgets or personalized emails loses customer trust without a customer service playbook dedicated to bad suggestions.
The e-commerce irrelevant recommendations customer support covers PDP widget, shopping cart, CRM email, and AI assistant complaints, structured feedback collection, empathetic response, targeted opt-out, and loop-back to merchandising, distinct from the implementation of recommendation engines.
This guide #439 covers IRECO-SUP policy, IRECO-FLOW workflow, and ireco KPI. First content on bad recommendations from the customer side. Distinct from contextual recommendations and assistant vs reco (#17): here, customer service playbook when the customer says the suggestion is bad: trust, feedback, and correction.
Summary
Why do bad recommendations generate support tickets?
An irrelevant recommendation ticket relates to a customer disputing a product suggestion (widget, email, push, bot) perceived as off-topic, repetitive, intrusive, or offensive, not a general product question.
Five typical customer pain points
Wrong category: men's recommendation for a female customer, baby product without children
Already purchased: same SKU suggested post-order
Cart duplicate: item already in the cart suggested as cross-sell
Spam recommendation email: too frequent, never relevant
Absurd AI Bot: assistant suggests incompatible or absurd options
Salesforce observes that recommendation clicks represent ~7% of traffic but ~24% of orders when relevant (Salesforce via Best for Ecommerce 2026). The reverse is true: irrelevant recommendations erode NPS and increase email unsubscribes. McKinsey estimates that 71% of consumers expect relevant personalization and react negatively to "fake personalized" (McKinsey, personalization 2026).
Angle #439 vs related content
Contextual recommendations: profile cart page engine setup. The #439 = customer support complaint post-display.
Assistant vs recommendation #17: lever strategic choice. The #439 = handling "your recommendation is bad".
Bot bad recommendation #440: future AI correct and learn. The #439 = agent playbook + merch loop.
Retention complaints: general service recovery. The #439 = personalization trust recommendation angle.
AI Governance #142: saying vs doing bot. The #439 = customer ticket bad bot suggestion.
DTC Example
Skincare DTC, widgets + Klaviyo post-purchase recommendations, 180 ireco tickets/year. Without IRECO-SUP: ireco_unsubscribe_spike 12%, NPS -6 pts complaint cohort. After playbook: ireco_ticket_rate -41%, ireco_feedback_to_merch 78% routed, ireco_csat 4.0/5 post-response.
Trust vs conversion
An agent minimizing it as "it's just an algorithm" worsens the sentiment. IRECO-SUP = recognize, explain without jargon, act (opt-out, feedback, alternative).
Post-purchase email peak ireco
Complementary flow D+3 post-order: highest ireco_already_owned if exclude rule lag. Audit Klaviyo exclude purchased weekly.
How does a recommendation complaint differ from a product question or marketing spam?
Complaint recommendation, product question PDP and newsletter unsubscription: three distinct support intents.
Matrix intent → dominant ticket
Bad reco #439: widget/email/bot suggestion perceived as incorrect
Product question: PDP size compatibility usage
Newsletter only: general unsubscription
Purchase objection: pre-purchase conversion barrier
Four IRECO-SOURCE
ireco_widget_pdp: related/complementary product detail page carousel
ireco_widget_cart: checkout cart drawer cross-sell
ireco_email_crm: Klaviyo post-purchase browse abandonment reco
ireco_bot_assistant: shopping assistant conversation suggestion
Reco + support stack
Shopify Search & Discovery, Nosto, Rebuy, Klaviyo flows, Gorgias tags ireco_*, Notion feedback merch queue, granular opt-out reco emails.
Promise #439
IRECO-SUP policy, IRECO-MAP matrix, 12 ireco_* typologies, IRECO-FLOW flow, IRECO-* macros, KPI ireco_* + merch loop.
Fake personalization
Customer sees first name + wrong product = worse than generic. IRECO-SUP handles this P2 case with enhanced empathy.
Shopify complementary recommendations
Shopify docs related vs complementary types map to ireco_widget_pdp IRECO-SOURCE for agent accuracy.
Which ireco_* typologies should be mapped?
Twelve irrelevant recommendation ticket typologies for consistent routing.
Twelve ireco scenarios
ireco_wrong_category: incompatible category, gender, age, or usage
ireco_already_owned: recently purchased product re-offered
ireco_duplicate_cart: same SKU already in cart widget
ireco_wrong_size_variant: reco size inconsistent with profile
ireco_price_tone_deaf: premium upsell out of budget signal
ireco_oos_in_widget: out of stock displayed in reco
ireco_allergy_skin_conflict: ingredient contraindicated for profile
ireco_email_frequency: too many irrelevant reco emails
ireco_email_unsubscribe: stop only reco, not all marketing
ireco_bot_bad_suggestion: AI assistant absurd suggestion
ireco_offensive_inappropriate: culturally inappropriate suggestion
ireco_feedback_how: how to report bad reco
Helpdesk tags
ireco, ireco_widget, ireco_email, ireco_bot, ireco_merch_flag, ireco_resolved. Distinct from product_question, newsletter_unsub.
Prioritization
P1: ireco_offensive_inappropriate, ireco_allergy_skin_conflict. P2: ireco_already_owned, ireco_wrong_category. P3: ireco_feedback_how FAQ, ireco_email_frequency.
Mining ireco verbatims
Export 90 days of "recommendation", "not relevant", "why do you suggest", "you might also like", "algorithm", "nonsense". Tag source widget vs email.
Which IRECO-MAP matrix should be documented?
The IRECO-MAP recommendations matrix lists sources, authorized responses, opt-out and merch escalation.
IRECO-MAP Columns
ireco_source: widget_pdp, widget_cart, email_crm, bot_assistant
engine_vendor: Shopify native, Nosto, Rebuy, Klaviyo, custom
response_template: IRECO-* macro by typology
opt_out_path: reco emails only, all marketing, widget dismiss
merch_escalation: yes/no, SLA 5 days fix rules
geste_policy: none, -10% next order, free ship once if P1 offensive
feedback_form_url: /pages/reco-feedback structured
exclude_rules_known: purchased 90 days, cart duplicate, OOS hide
Example ireco_email_crm post-purchase
ireco_email_crm: Klaviyo flow complementary, opt_out_path reco segment only, merch_escalation yes SKU flagged, exclude_rules purchased 60 days should apply document gap if failing.
Example ireco_widget_pdp related
ireco_widget_pdp: Shopify related products, response IRECO-WIDGET-01, opt_out none site-wide but feedback form, merch_escalation if wrong_category > 5 tickets same PDP/month.
Publication /pages/reco-feedback
FAQ IRECO-MAP: why these suggestions, how to report, disable reco emails, difference reco vs newsletter.
Merch queue fields
ireco_merch_ticket: source, SKU shown, SKU context page/cart, customer segment, verbatim, suggested fix exclude rule.
Nosto Rebuy vendor field
IRECO-MAP engine_vendor per widget documents honest agent explain "Nosto third-party engine" vs Shopify native.
How to draft the IRECO-SUP policy in eight rules?
The IRECO-SUP irrelevant recommendation policy governs empathy, customer action, and the feedback loop.
Eight IRECO-SUP Rules
Acknowledge first: acknowledge frustration, do not downplay "algorithm"
Identify ireco_source: widget, email, bot before macro
No blame customer: never "you navigated incorrectly"
Granular opt-out: email recommendations vs all marketing per IRECO-MAP
Flag merch IF-7: ireco_merch_flag if repeat SKU or wrong_category pattern
Bot bad suggestion: log conversation_id route_#440 corpus + handoff if harmful
Geste policy bounded: geste_policy only for offensive P1 or 3+ ireco same customer within 90 days
Close with feedback CTA: /pages/reco-feedback link on every resolved ireco
Already owned response
Verify order history SF-4. If purchased < 90 days → IRECO-OWNED-01 apologize + confirm exclude rule should apply + merch flag if engine failed.
Email unsubscribe reco only
Klaviyo segment "product recommendations" unsubscribe, not global marketing unless client asks for all.
Offensive P1
ireco_offensive_inappropriate: manager review within 24 hours, merch kill rule, compensation if policy, personal response not template only.
How to apply the IRECO-FLOW process in eight steps?
The IRECO-FLOW framework structures the processing of irrelevant recommendation tickets.
Eight steps IF-1 to IF-8
IF-1 Intake: verbatim, chat email social channel
IF-2 Classifier ireco_*: section 3 typology
IF-3 Match IRECO-MAP: source engine opt_out action
IF-4 Verify context: order history, cart snapshot, email flow name, bot log
IF-5 Respond: IRECO-* macro grounded in IF-4
IF-6 Client action: opt-out, feedback form, alternative product manual if policy
IF-7 Merch flag: Notion ticket SKU rule gap if pattern
IF-8 Document: ireco_type, source, SKU flagged, opt_out_done
IF-4 ireco_already_owned
Shopify orders API last 90 days SKU match reco shown. If match → IF-7 merch exclude purchased rule failure. IF-5 IRECO-OWNED-01.
IF-6 ireco_email_unsubscribe
Execute Klaviyo reco segment unsub IF-6. Confirm email IF-5 IRECO-UNSUB-01. Not global unless requested.
IF-4 ireco_bot_bad_suggestion
Pull bot transcript session_id. Attach IF-8. Route harmful allergy conflict to P1. Sync #440 corpus feedback.
IF-7 wrong_category pattern
Same PDP widget > 5 ireco_wrong_category 30 days → merch audit collection rules tags gender age.
IF-6 alternative manual
Optional white-glove: agent suggests 1-2 SKU grounded in catalog if client asked « what do you really recommend » post-complaint. No random upsell.
IF-8 export to #440 bot corpus
ireco_bot_bad_suggestion transcripts JSONL monthly sync future bot correction training set.
Which IRECO-* macros and touchpoints should be configured?
Eight irrelevant agent recommendation macros and feedback touchpoints.
IRECO-ACK-01 (acknowledgement)
"You are right to point out: [suggested product] does not match [customer context order/cart]. Thank you, this helps us improve our suggestions."
IRECO-OWNED-01 (already purchased)
"This product was in your order #[X] of [date]. It should no longer be recommended: we are correcting the exclusion rule. Sorry for the noise."
IRECO-WIDGET-01 (PDP/cart widget)
"The "you might also like" blocks are automatic. Your report has been forwarded to the product team. Report here: [feedback_form_url]."
IRECO-UNSUB-01 (stop reco emails)
"You have been unsubscribed from product recommendation emails. You will still receive orders and promotions if you are subscribed to global marketing."
IRECO-BOT-01 (AI assistant)
"The chatbot suggestion of [date] was unsuitable. Conversation forwarded to the AI team. Would you prefer to speak to an advisor regarding [initial need]?"
IRECO-OOS-01 (out of stock in reco)
"Recommended product out of stock: display error. Corrected on the catalog side. Stock alternative: [SKU link] if you wish."
Touchpoints
Footer widget "Irrelevant suggestion?" → feedback form
Email footer "Customize my recommendations"
/pages/reco-feedback structured form
Help center article IRECO-MAP FAQ
Post-chat CSAT tag ireco if reco complaint
IRECO-FREQ-01 (too many emails)
"Reco frequency reduced to max 1/week. Segment updated. Full reco opt-out: [link]."
CSAT post-ireco resolution
Auto-send CSAT tag ireco 24 h post close. Feed ireco_csat KPI and macro A/B IRECO-ACK variants.
Which AI bot, allergy, and merchandising loop use cases should be addressed?
Special ireco cases require IRECO-MAP extensions and separate SLAs.
Bot assistant #440 overlap
ireco_bot_bad_suggestion : IF-8 log → corpus bot #440. Link AI governance (#142) P2 incident if repeat intent. Agent IF-5 IRECO-BOT-01 not re-argue bot right.
Skincare allergy P1
ireco_allergy_skin_conflict : if reco contains allergen client profile stated → P1 merch kill + gesture if policy + confirm not medical advice.
Gender age wrong_category
Merch audit tags gender collection rules. IF-7 mandatory if > 3 reports same widget.
Content gaps merch
Repeat ireco on PDP with thin tags → info gaps (#173) merch + support joint fix.
Complaints retention overlap
Client threatens churn over reco spam → service recovery retention + IRECO-UNSUB-01 IF-6.
Routine reco bot #391 adjacent
Bad qty or bundle reco distinct ireco_wrong_category. Route bot quantity (#391) if complaint is format not product type.
Shopify related API limits
Native related sometimes weak. Document IRECO-MAP engine_vendor per placement so agent explains source honestly.
Social DM ireco complaints
Instagram screenshot bad reco email : IF-1 channel social, same IRECO-FLOW, public reply take offline template.
Which ireco KPIs should be measured?
Irrelevant recommendation support KPIs drive trust, opt-out, and the merchandising loop.
Eight key metrics
ireco_ticket_rate: ireco tickets / orders or sessions with reco
ireco_fcr: first contact resolution / ireco tickets
ireco_feedback_form_rate: forms submitted / ireco tickets
ireco_merch_flag_rate: IF-7 flags / ireco tickets
ireco_merch_fix_sla: rules fixed within 5 days / flags
ireco_reco_unsub_rate: reco segment unsub / ireco email tickets
ireco_repeat_complainer_rate: 2+ ireco same email 90 days
ireco_csat: ireco tag post-resolution satisfaction
DTC Benchmark
ireco_ticket_rate < 0.3% orders with reco widgets, ireco_fcr > 72%, merch_fix_sla > 80%, repeat_complainer < 8%, ireco_csat > 3.9/5.
Monthly dashboard
IRECO-SOURCE breakdown, top SKU flagged, top PDP widgets, ireco_type distribution, correlation reco unsub vs NPS.
Merch product review
Monthly 30-min support + merch: review IF-7 queue, close loop, update IRECO-MAP exclude_rules_known.
Widget CTR vs ireco tickets
High CTR widget + high ireco_wrong_category same placement = rules broken not customer wrong.
NPS cohort ireco complainers
Track NPS of customers with ireco ticket vs control. Target recovery +5 pts post IRECO-ACK within 7 days.
Which anti-patterns should be avoided on recommendation complaints?
Ten anti-patterns for irrelevant support recommendations to ban.
1. "It's normal, algorithm"
Rule 1 acknowledge. Minimize destroys trust McKinsey fake personalized.
2. Blame client navigation
Rule 3 forbidden. Even if browse history caused reco explain gently.
3. Global unsub when reco only asked
Rule 4 granular. Client loses order emails accidentally.
4. No merch flag ever
Rule 5 IF-7. Same bug repeats 100 tickets.
5. Defend bad bot suggestion
Rule 6 IRECO-BOT-01. Agent argues bot was right = CSAT crash.
6. Random upsell in apology
IF-6 alternative only if client asks. Apology message not sales pitch.
7. Ignore allergy conflict
P1 SLA. Legal and health reputation risk.
8. No feedback form CTA
Rule 8 every close. Structured data beats chat only.
9. Confusing product question
IF-2 "does this product fit?" ≠ ireco "why offer X". Router product bot.
10. Gesture every ireco
Rule 7 bounded. Trains serial complainers.
11. OOS widget no merch ticket
ireco_oos_in_widget always IF-7 catalog hide rule.
12. Silo support vs marketing
Klaviyo owner not in monthly ireco review = repeat email ireco.
How does Qstomy help with irrelevant recommendations?
Qstomy on Shopify : IRECO-FLOW classify ireco_*, order history verify owned SKU, IRECO-ACK-01 and IRECO-UNSUB-01 templates, bot bad suggestion log handoff, feedback form CTA, merch flag pre-filled fields IF-8.
ireco Qstomy Capabilities
ireco_classify : IF-2 typology 12 intents
ireco_order_lookup : IF-4 already_owned verify
ireco_map_explain : source engine opt_out cite
ireco_ack_template : IRECO-ACK-01 auto
ireco_bot_log_attach : session transcript IF-8
ireco_merch_flag_export : IF-7 Notion webhook
Pipeline #439 → #440
#439 CS agents trust feedback merch. #440 future bot correct learn suggestion. Shared IRECO-MAP feedback corpus.
Encrypted DTC Scenario
180 ireco tickets/year baseline.
After IRECO-SUP + Qstomy : ireco_ticket_rate -38 % (preventive FAQ widget), ireco_fcr 79%, ireco_merch_fix_sla 85%, ireco_csat 4.1/5.
Explore customer support and request a demo.
Sales assistant alignment
See sales assistant for bot suggestion quality loop with IRECO feedback corpus.
What is the checklist for deploying IRECO-SUP?
IRECO-SUP Checklist (12 steps)
Inventory active IRECO-SOURCEs (PDP, cart, email, bot)
Document IRECO-MAP engine opt_out merch action per source
Draft IRECO-SUP policy 8 rules
Publish /pages/reco-feedback form + FAQ
Footer widget "not relevant" feedback link
Klaviyo reco segment granular unsub
Create IRECO-* helpdesk macros
Train agents IRECO-FLOW 45 min (IF-4 order verify, IF-7 merch flag)
Notion merch queue IF-7 SLA 5 days
Monthly support+merch ireco review 30 min
ireco_* tags + dashboard KPI section 9
Sync IRECO feedback → bot #440 future corpus
At a glance
#439 = customer reco complaint, not reco engine setup
IRECO-MAP: source → response → opt-out → merch
IRECO-FLOW: classify → verify → ack → action → flag
Acknowledge first: never downplay the algorithm
KPI ireco_merch_fix_sla: closed loop 5 days
FAQ
Difference with contextual recos?
Conversion engine setup. #439 = customer support agent gets a complaint about a bad suggestion.
Customer only wants to stop reco emails?
IRECO-UNSUB-01 Klaviyo reco segment, not global marketing.
Already purchased product suggested again?
IF-4 order verify + IRECO-OWNED-01 + IF-7 exclude rule failure.
Did the bot suggest nonsense?
IRECO-BOT-01 + log #440 corpus. Do not defend the bot.
How to report?
/pages/reco-feedback + footer widget link for every IRECO-ACK closed.
Go further
This week: publish IRECO-MAP /pages/reco-feedback, add feedback link to footer widget, create macros IRECO-ACK-01 and IRECO-UNSUB-01, schedule monthly IF-7 merch review.
Share this guide #439 with support and marketing: a sincere apology + granular opt-out + merch flag is worth ten promo codes, an "it's the algorithm" response is worth a global unsubscribe and a 2-star Trustpilot review.
Offensive reco or allergy?
P1 SLA 24 hrs manager + merch kill rule + bounded gesture IRECO-MAP.
Enzo
July 1, 2026
