From Campaign Calendar to Customer Memory

The obvious way to run CRM for an electronics retailer is to copy an e-commerce playbook: browse abandonment, cart abandonment, sale reminders, coupons, repeat.

I think that is the wrong starting point for Croma.

Electronics is a high-consideration, high-value and naturally low-frequency category. Someone may research a ₹70,000 television for a week and then not need another TV for years. A laptop buyer can still become relevant for a mouse, bag, extended protection, software support or an eventual upgrade. An air-conditioner purchase creates installation and maintenance moments long after checkout. And a customer who buys in a physical Croma store should not become a stranger when they open the app later.

Croma itself describes an unusually broad omnichannel footprint: 20,000+ products, 500+ brands, 540 Croma stores and 16 TRiBE stores across 200+ cities, with shopping across stores, Croma.com and Tata Neu.[1] It also operates post-purchase services through ZipCare and a loyalty layer through Croma First/NeuCoins.[2][3]

That creates a better CRM opportunity than simply sending more campaigns.

The CRM system should answer six questions for every reachable customer:

1. Who is this person?
2. What are they considering right now?
3. What do they already own?
4. What is the next useful moment in that product’s lifecycle?
5. Which channel is appropriate for that moment?
6. What should we deliberately not send?

The goal is not maximum communication. The goal is maximum incremental customer value per useful interaction.

Croma’s CRM should not behave like a campaign calendar. It should behave like customer memory: what someone is considering, what they already own, what needs attention next, and when Croma has a genuinely useful reason to show up.

A grocery app can reasonably expect weekly purchase behaviour. Fashion can create newness every few days. Consumer electronics behaves differently.

Replacement cycles vary dramatically by category. Purchase intent is often concentrated into short windows. The offline store matters. Service matters. Financing, exchange, protection and installation can matter almost as much as the product page.

That changes how I would judge CRM.

A customer who bought a refrigerator six months ago and has not opened the Croma app since is not necessarily “churned”. They may be completely satisfied. Sending a generic We miss you coupon misunderstands the relationship.

Likewise, a customer who has viewed the same laptop four times in two days may be far more commercially relevant than someone who opened the app ten times this month without meaningful intent.

So I would design Croma CRM around four principles:

1. Intent beats raw recency

A high-intent product action should matter more than an arbitrary “active/inactive” label.

2. Ownership matters after checkout

The purchase event should start the next lifecycle, not end the journey.

3. Utility earns the right to promote

Delivery, setup, protection, service, price and availability messages create trust. Promotional communication should sit on top of that relationship, not replace it.

4. Incrementality beats engagement metrics

Open rate and CTR help diagnose creative and delivery. They should not be confused with business impact.

Croma is not a single-channel retailer. A customer can research online, walk into a store, purchase offline, activate benefits through Tata Neu and return later through another surface.[1][2]

That means the first question I would ask is not:

Which campaign should we launch?

It is:

Can we reliably recognise the same customer across these moments?

If an in-store television purchase does not update the same profile used for digital campaigns, CRM can immediately become annoying: the customer keeps receiving TV acquisition messages after they already bought one.

For me, identity resolution and event quality are part of CRM execution, not somebody else’s backend problem.

The minimum customer memory I would want available

Identity and reachability

• Customer ID / unified mobile number
• Push reachability
• WhatsApp consent/reachability
• Email reachability
• SMS reachability
• Preferred language, where available
• City / serviceable location
• Preferred or most-used store, where available

Commerce

• Last purchase date
• Purchase channel: store / web / app / Tata Neu, where available
• Product category and SKU
• Order value
• Brand
• Payment state
• Exchange usage
• Croma First / loyalty status, where available

Ownership

• Product owned
• Delivery date
• Installation status, if applicable
• Protection/service plan status
• Warranty or plan eligibility window, where available
• Service history

Behaviour

• Search
• Product view
• Product comparison
• Add to cart
• Checkout started
• Payment failed
• Price-drop or back-in-stock interest
• Category affinity

MoEngage supports segmentation using user attributes, events and event attributes, and can orchestrate communication across channels including push, email, SMS and WhatsApp.[4] Real-time segment evaluation can also support use cases such as abandoned carts and in-store purchase follow-up when the underlying events are available.[5]

The platform is capable. The harder work is deciding what deserves to become a trigger.

Campaign calendars are useful operationally. They are a poor customer model.

A customer does not experience “Diwali Campaign 07”. They experience one continuous relationship with Croma.

I would organise the CRM architecture into six lifecycle states:

This sounds simple. That is a feature, not a bug. If the lifecycle model cannot be understood quickly by CRM, product, analytics and business teams, it will be difficult to operate consistently.

I would not start by building 40 flows. I would start with the moments where intent is clear and where Croma has something useful to offer.

Journey 1: High-intent browse and cart recovery

Trigger

A user views the same product repeatedly, compares products, adds an item to cart, or begins checkout but does not purchase.

Logic

The message should reflect the strength of intent.

A casual product view does not deserve a WhatsApp message. A ₹60,000 cart abandonment probably deserves more attention than a single browse of a ₹799 accessory.

A possible flow:

1. 1

   Product viewed repeatedly

2. 2

   Wait 1–2 hours

3. 3

   If purchased, exit

4. 4

   Push with the exact product, availability, or a useful reason to return

5. 5

   If a high-intent cart remains open, consider WhatsApp

6. 6

   If the research cycle is longer, email a comparison or category guidance

7. 7

   Exit immediately on purchase

How I would build it in MoEngage: use an event-triggered Flow, let real-time segment evaluation qualify the high-intent action, use a wait/condition stage before the next channel, and set purchase as the conversion/exit condition. Frequency capping should still apply so one aggressive recovery journey does not collide with the rest of the CRM calendar.[5][9][10]

I would not default to a discount. First test whether price, stock, delivery speed, exchange value, financing, product comparison or simple continuity is the actual blocker.

Primary KPI

Incremental purchase conversion versus a holdout/control cohort where sample size allows.

Guardrails

Opt-outs, WhatsApp blocks, push disablement, margin dilution and message frequency.

Journey 2: Payment-failure rescue

A payment failure is one of the clearest intent signals in commerce. The customer was trying to buy.

Trigger

payment_failed with no successful purchase within a defined window.

Channel hierarchy

• Push when reachable
• WhatsApp for high-intent recovery where consent exists
• SMS as a fallback for critical reach, where appropriate

The copy should be functional, not promotional:

Your payment did not go through. Your cart is still available if you’d like to continue.

No unnecessary urgency. No fake scarcity. No exposing payment information in the message.

KPI

Recovered orders and recovered contribution margin, not click-through rate.

Journey 3: Delivery-to-ownership onboarding

Most CRM teams celebrate the purchase event. For Croma, that is where a valuable new dataset begins.

Trigger

delivery_completed

Example sequence

• Day 0: Delivery confirmation, plus the next step if installation is required
• Day 1–3: Setup and help content relevant to the product category
• Eligibility window: ZipCare or protection communication if applicable
• Later: Relevant accessories or complements
• After sufficient usage: Review, NPS, or service feedback

Croma’s ZipCare proposition spans setup, protection, maintenance and repair services across more than 40 gadget and appliance categories.[3] That makes ownership-stage CRM a real commercial and customer-experience opportunity, not filler communication.

I would only promote an accessory when it makes sense for what the person bought. A TV buyer may need a soundbar or wall-mount-related service. A laptop buyer may need a bag, mouse or protection. An air-conditioner buyer needs a very different lifecycle.

The product category should shape the journey.

Journey 4: Protection, warranty and maintenance lifecycle

This is where Croma can do something many e-commerce CRM programs cannot: remain useful long after the purchase.

Croma states that some ZipCare protection plans can be purchased after the device purchase, while maintenance plans have their own eligibility windows and can apply to products bought online or offline.[3]

That means timing matters.

Example triggers

• Protection plan eligible but not purchased
• Eligibility window approaching closure
• Manufacturer warranty approaching expiry, where data is available
• Seasonal maintenance window for relevant appliances
• Service request completed

Example

A customer who bought an AC does not need another AC campaign two months later. They may need installation help, maintenance at the right seasonal moment, or service support.

That is what I mean by customer memory.

KPI

Protection/maintenance attach rate, service conversion and downstream customer value.

Journey 5: Cross-category household expansion

Traditional cross-sell often means “people who bought X also bought Y”.

That is a start, but Croma can go further because the customer’s purchase history begins to describe the technology setup of a household.

Examples:

• Smartphone → earbuds / smartwatch / charger ecosystem
• Laptop → mouse / keyboard / storage / bag / monitor
• TV → soundbar / streaming accessory / gaming
• Camera → storage / tripod / accessories
• Home appliance → maintenance or adjacent home category when intent appears

The key phrase is when intent appears.

I would combine ownership data with fresh behaviour instead of blasting every possible accessory after purchase.

A customer who bought a MacBook three months ago and suddenly starts viewing monitors is more interesting than a generic “MacBook owner” segment.

KPI

Incremental cross-category revenue, margin and category penetration.

Journey 6: Upgrade and replacement

“We miss you” is rarely a useful electronics strategy.

A better question is:

Is this customer entering a realistic replacement window for something they already own?

The timing should be category-specific and validated against actual Croma purchase data.

Once an ownership-age signal combines with fresh product interest, the CRM can become much more useful:

• Exchange availability
• Upgrade comparison
• New-generation product education
• Store consultation
• Price-drop alerts
• Financing options, if relevant and compliant

Croma supports product exchange for eligible same-category purchases, creating a natural CRM lever when replacement intent appears.[6]

The important thing is not to invent a universal replacement cycle. A smartphone, refrigerator and television should not share the same rule.

One of the easiest ways to damage CRM is to use the strongest channel for everything.

Croma already uses WhatsApp as a service surface: its public FAQ points customers there for invoices, order tracking, stock checks and store discovery.[8] That makes the channel more valuable, not less. I would be careful not to dilute that trust by turning the same surface into an indiscriminate promotional pipe.

I would establish a channel hierarchy based on urgency, intent, richness, reachability and customer preference. MoEngage exposes channel-level reachability, so the orchestration should use what the platform actually knows about whether a customer can be reached rather than assuming every customer is equally reachable everywhere.[11]

MoEngage’s retail stack is explicitly designed for cross-channel engagement and supports push, email, SMS, WhatsApp, in-app and web experiences.[7]

But orchestration is more than adding channels to the same flow.

I would also define message priority

1. Transactional / service-critical
2. Customer support / ownership
3. High-intent lifecycle
4. Loyalty / benefit
5. Promotional campaign

If someone has a delivery or service issue, I do not want a sale blast landing on top of it because two teams happened to schedule campaigns independently.

CRM should coordinate the relationship, not expose the org chart.

RFM is a good framework for identifying valuable, recent or dormant customers. MoEngage supports RFM-style and behavioural segmentation, and retail brands use it to create more relevant journeys.[4][7]

For Croma, I would not stop there.

A single RFM model can be misleading in electronics because expected frequency differs by category.

Someone who bought a ₹1 lakh refrigerator 18 months ago may have more lifetime value than someone who purchased three low-ticket accessories last quarter.

I would layer segmentation across four dimensions:

Value

• High historical value
• High potential value
• Low-value / deal-led
• Service/protection customer

Intent

• Browsing
• Repeated product interest
• Cart
• Checkout
• Payment failure
• Price/stock watcher

Ownership

• Category owned
• Product age
• Protection status
• Service history
• Upgrade eligibility

Relationship

• New customer
• Repeat customer
• Cross-category customer
• Croma First / loyalty member
• Store-led vs digital-led
• Dormant with no current intent

The intersection matters more than the label.

“High-value + laptop owner + monitor browsing + WhatsApp reachable” is actionable.

“Premium customer” on its own is not.

The role I am solving for is execution-heavy. A beautiful lifecycle map is useless if campaigns break at launch.

Before any meaningful campaign or flow goes live, I would expect a standard QA checklist.

Pre-launch CRM QA

Audience

• Segment logic verified
• Audience count sanity-checked
• Exclusions and suppression lists applied
• Purchase/event exit conditions confirmed
• Channel reachability checked

Message

• Personalisation variables tested
• Fallback values present
• Product names/prices pulled correctly
• Copy approved
• WhatsApp/SMS template status confirmed
• Language version checked where applicable

Experience

• Deep link tested on Android/iOS/web as relevant
• Landing page/product page available
• Promo code or benefit works
• Expired/out-of-stock states handled

Measurement

• UTM/tracking parameters present
• Campaign naming/tagging follows the agreed taxonomy
• Conversion event verified
• Holdout/control configured where appropriate
• Dashboard/reporting owner defined

Delivery and compliance

• Frequency cap checked
• Send time reviewed
• Consent, template and applicable regulatory requirements checked for the channel
• Test devices/accounts used
• Duplicate journey exposure checked
• Exit criteria tested

I would rather delay a campaign than send a personalised message with the wrong product, broken deep link or an offer the customer cannot redeem.

That is not operational perfectionism. That is brand protection.

The operating system behind the journeys

The role is not only about designing journeys. Someone has to run the machine every day.

I would keep one lightweight campaign operating standard across MoEngage and reporting:

Brief → Build → QA → Approval → Test send → Schedule → Live check → Readout

Every campaign or Flow should have a clear owner, objective, audience, exclusions, channel, conversion event, start/end condition and measurement plan before it is scheduled. The same naming convention should carry through MoEngage, UTMs and dashboards so a campaign can be traced without detective work later.

For higher-risk sends such as large WhatsApp or SMS campaigns, I would use a second-person check before launch. After launch, I would watch delivery/errors, audience volume, deep links and conversion-event firing early enough that a bad send can still be stopped rather than explained the next morning.

A campaign calendar is useful. But it should be the output of this operating system, not the strategy itself.

A/B testing is often reduced to subject lines and button colours.

Those tests are fine. They are not where I would spend most of the learning budget.

The bigger questions are:

Does the message need an incentive?

Cart recovery with useful information vs cart recovery with a discount.

Which channel earns the conversion?

Push only vs push followed by WhatsApp for high-intent users.

When is the next message actually useful?

Accessory recommendation on Day 1 vs Day 7 vs after fresh accessory browsing.

Can we reduce sends without reducing sales?

Broad campaign audience vs intent-filtered audience.

Does service-led CRM improve future commerce?

Ownership/service journey exposed vs matched holdout.

Where scale allows, I would keep persistent or journey-level holdouts so the CRM team can answer the uncomfortable question:

Would these customers have purchased anyway?

Without that, attribution will flatter us.

I would not run this program on open rates.

Primary business metrics

• Incremental orders from CRM journeys
• Incremental revenue / contribution margin
• Conversion rate by lifecycle journey
• Cross-category purchase rate
• Repeat purchase rate within category-appropriate windows
• ZipCare/protection/maintenance attach rate where relevant
• Loyalty activation / usage where relevant

Funnel metrics

• Product view → cart
• Cart → checkout
• Checkout → purchase
• Payment failure → recovered purchase
• Delivery → ownership action
• Eligibility → protection/service conversion

Channel health

• Reachability by push / WhatsApp / email / SMS
• Delivery rate
• Opt-out / unsubscribe / block rate
• Push disablement
• Template rejection/error rate
• Send frequency per customer

Diagnostic engagement

• Open rate
• Click-through rate
• Session return

Useful? Yes.

North star? No.

Days 1–30: Understand the machine before changing it

I would audit:

• Current MoEngage workspace structure
• Events and user attributes actually available
• Identity stitching between store and digital purchases
• Existing flows and campaigns
• Overlapping journeys
• Frequency caps and suppression rules
• Push, WhatsApp, email and SMS reachability
• Delivery, template and campaign failure rates
• Frequency-cap, DND/preference and suppression behaviour
• Campaign naming, approval and QA process
• Existing reporting and attribution
• Top customer complaints related to communication
• Conversion volume at the main funnel drop-offs

The output would be a simple lifecycle coverage map: what exists, what is broken, what is missing, what should be retired.

Days 31–60: Fix the highest-intent journeys

I would prioritise three journeys first:

1. Payment failure recovery
2. High-intent browse/cart recovery
3. Delivery-to-ownership onboarding

For each, I would establish clean exits, channel logic, QA standards, dashboards and a holdout where feasible.

This gives the team measurable wins without trying to redesign the entire CRM program at once.

Days 61–90: Build the ownership engine

Then I would move into the part that can differentiate Croma:

• ZipCare/protection eligibility
• Warranty/maintenance moments
• Category-specific accessory journeys
• Cross-category expansion
• Replacement/upgrade signals
• Loyalty/Croma First activation where applicable

By Day 90, I would want fewer disconnected campaigns and a clearer set of always-on customer journeys.

Imagine I buy an air conditioner from a Croma store.

A weak CRM system knows:

Saurav bought something 14 days ago.

A better CRM system knows:

Saurav bought an AC offline. Delivery is complete. Installation status is known. He does not need another AC acquisition ad. He may be eligible for a relevant protection or maintenance plan. When seasonal servicing becomes useful, that moment can be triggered. If he later starts browsing air purifiers or smart-home products, that is fresh intent, not an arbitrary cross-sell calendar.

The second system feels less like marketing.

It also creates more opportunities to make money.

That is the point.

I would not make WhatsApp the default channel

High reach is not permission to send low-value messages.

I would not discount every abandoned cart

A discount can train customers to abandon. First diagnose the friction.

I would not call every inactive customer “churned”

Electronics has long natural purchase cycles.

I would not treat an offline buyer as digitally anonymous

If identity stitching is available, store purchases must inform digital communication.

I would not use one replacement rule for every category

Product lifecycles differ too much.

I would not optimise the program for opens and clicks

The business pays for incremental customer value, not dashboard activity.

And I would not launch automation just because MoEngage makes it possible

Automation scales good decisions and bad ones equally well.

Croma does not need CRM to make customers buy electronics every week. That would be the wrong ambition.

It needs CRM to be exceptionally good at recognising the relatively small number of moments when a customer genuinely needs something:

• A product they are seriously considering.
• A payment that failed.
• A delivery that needs setup.
• A device that needs protection or maintenance.
• An accessory that makes the purchase more useful.
• A product approaching a sensible upgrade window.
• A new category need that has actually appeared.

MoEngage can orchestrate those moments across push, WhatsApp, email and SMS. The harder and more valuable work is the judgment underneath it: event quality, segmentation, channel choice, QA, experimentation, suppression and measurement.

My operating principle would be simple:

Send fewer messages that know more about the customer.

That is how CRM moves from a campaign function to a customer-value engine.

How should an electronics retailer like Croma use CRM?

Electronics CRM should be organised around purchase intent and product ownership rather than daily engagement. The highest-value journeys typically include high-intent browse/cart recovery, payment recovery, delivery/setup, protection and maintenance, relevant accessories, loyalty and eventual upgrade/replacement moments.

How can MoEngage be used for Croma-style retail CRM?

MoEngage supports event- and attribute-based segmentation plus cross-channel journeys across push, email, SMS and WhatsApp. A retailer can use those capabilities to trigger communications from actions such as repeated product views, cart activity, payment failure, purchase, delivery, service eligibility and renewed category intent, assuming those events are correctly integrated.[4][5]

Which CRM channel should be used for what?

Push works well for lightweight real-time nudges; WhatsApp should be reserved for higher-intent or genuinely useful service moments; email is better for richer comparisons and ownership content; SMS is valuable for concise critical or fallback communication. The right mix depends on consent, reachability, urgency and measured customer response.

What should be the north-star metric for retail CRM?

Not open rate. The most useful business measure is incremental customer value: additional orders, revenue or contribution margin caused by the CRM intervention versus what would have happened without it. Engagement metrics are diagnostics, not the final outcome.

Saurav Mishra works at the intersection of growth, CRM/lifecycle, product and analytics. He currently leads growth at Xpat, where his work spans lifecycle journeys, segmentation, retention, acquisition, experimentation and customer communication across channels including push, WhatsApp and email. Earlier, he worked in product analytics and co-founded a growth agency.

This case study is part of a collection of independent breakdowns and operating ideas on growth, CRM and product.

Read about Saurav Mishra, explore other growth case studies, or get in touch.

1. Croma — About Croma. Public company information including omnichannel presence, assortment, brands and store footprint.
2. Croma — Croma First / Loyalty Program Terms. Public information on Croma First and NeuCoin benefits.
3. Croma — ZipCare. Public information on setup, protection, maintenance and repair services, including product coverage and purchase windows. ZipCare, ZipCare Protect, ZipCare Maintain.
4. MoEngage — Basics of Segmentation. Documentation on grouping users by attributes, events, and event properties.
5. MoEngage — Real-Time Segment Evaluation. Documentation and retail/e-commerce use cases for real-time event-based segmentation.
6. Croma — Payment / Exchange FAQs. Public information on eligible product exchange mechanics.
7. MoEngage — Retail & E-commerce. Public overview of MoEngage cross-channel retail capabilities.
8. Croma — FAQ / Croma Assist. Public information on Croma’s WhatsApp service for invoices, order tracking, stock availability and store discovery.
9. MoEngage — Frequency Capping. Documentation on delivery controls and frequency capping across push, email, SMS, WhatsApp and Flows.
10. MoEngage — Create a Flow. Documentation on building a Flow, including entry, conversion and exit conditions, control groups, waits, and A/B testing.
11. MoEngage — User Reachability. Documentation on channel reachability across push, email, SMS and WhatsApp.