AI Marketplace - FAQs
Key Questions to ask before Launching an AI-led Offers Marketplace
In today's fast-paced business landscape, staying ahead of the curve is essential for success.
As a senior executive at a bank, you're always on the lookout for innovative solutions that can give
your organization a competitive edge. An AI-led offers marketplace that enables customers discover,
engage and redeem merchant offers could be the answer you've been searching for. But before diving
in, have you identified the needs of all stakeholders to ensure long-term success? Here are some
questions that you should consider asking.
How can customers easily discover available offers?
Ensure that an offers marketplace has features to improve discoverability like category
listings, campaign banners, natural language search, location pickers, Netflix-like
curated merchant lists. Ensure the marketplace is available in English and the next
most spoken language in the country.
How can the experience be personalized for every customer?
Ensure that the experience is personalized through personalized lists, personalized campaigns, similar recommendations
- Create cold start lists to personalize the experience when there is only digital interaction data available
- Create warm start lists to personalize the experience when transaction data is also available
- Provide a reasoning for every recommendation
- Enable customers to like/dislike the recommendation and mechanism to learn from this feedback to improve the recommendation quality.
Can customers access it on any channel?
Customers should be able to access offers on a PWA, on website, within mobile app and on any other channel like Email, SMS, WhatsApp that the bank is using today to communicate with customers.
Will customers have to visit an external link?
Marketplace should ideally be hosted on Bank’s domain. In such a case, customers will not have to go to an external link
Will the experience change with location?
Yes, the experience should change by location with recommended merchants being more relevant to the current location while accounting for customer’s preferences and needs. Customers should have the option of seeing these merchants on a map.
Can customers access it even while travelling abroad?
Yes, ideally the bank should have offered that customers can redeem even while travelling abroad.
Can fulfilment be made seamless for the customer?
Simplify fulfilment by reducing the steps in the process. Ensure that the following fulfilment journeys are supported:
- Online
- Simply using a coupon code on redirected website
- Bank specific codes in 90%+ cases.
- BIN locked where merchant has the capability.
- In-store
- Simply using the bank card.
- No need to enter any code, scan any QR etc. (BIN locked where POS has capability)
- No need to come back/open the app for repeat usage.
How can disputes (if any) be resolved easily?
Setup a dedicated customer services team to handle disputes.
- Level 1 contact for customers will need to be Bank’s customer support to safeguard PII data. If the bank is working with a vendor to source offers, then queries and issues can be passed from Bank team to vendor’s dedicated customer support for resolution.
- Discuss with merchant on what are the expected disputes and standard resolution mechanisms to ensure faster TAT and superior customer experience. For example, in case of a valid discount not being honoured, merchant should provide the customer a compensation of equal/larger value through store credit/voucher code/gift card etc.
How can the merchant offers be sourced?
Offers are usually sourced from offer aggregators, affiliate networks, card network schemes and from direct merchants
How many offers are needed every month?
The exact numbers will vary based on the size of the bank and their requirements. A mid-to-large-sized bank (500K - 1M customers) should ideally have 400-600 offers. The focus should not be on quantity but rather higher quality of offers.
What categories should merchants belong to?
Offers should cover categories like travel (holidays, flights, hotels, attractions), F&B, and Retail (clothing, electronics, groceries, etc.)
What kind of offers will be relevant to customer needs?
Offers linked to customer’s transaction behaviour tend to be relevant. Leverage customer’s transaction behaviour to create and constantly update an expansive, target list of merchants in the acquisition pipeline across categories:
- High Spends - where customers are spending today and are likely to spend more if incentivized (top merchants list).
- Potential Spends - merchants where a given customer cohort is not spending at all or spending less frequently, but likely to spend when incentivized. We use our algorithms for this.
- Market Movers - merchants that do not have large spends historically but are witnessing accelerated spends in recent times (MoM growth, QoQ growth)
- Market Research - merchants that are popular on internet platforms and are the talk of the town (using social media, trip advisor, blog posts etc.
- Seasonal events – merchants that suit event-based campaign like Ramadan, Back to School, National Day, Diwali etc
Should offers be long term or short term?
A combination of both short term and long-term offers is recommended. Long duration offers are generally annual, generic in nature and the discount cost is borne by the merchant in 99% of cases. Tactical offers are seasonal, short duration offers, typically ranging from a week to 2 months, exclusive to bank in most cases where the cost of the discount is generally shared between the merchant and bank.
How can the offers be best in class?
- Ensure some of the offers have exclusivity. The exclusivity varies based on the term over which offers are made available by merchants. Usually, 90% of short term/festive offers like Ramadan, Chinese New Year, Back to School will be exclusive and on an average 50% of long-term offers will be exclusive. Ensure that all offers are available to all customers regardless of the country they belong to.
- Conduct a periodic (bimonthly) monitoring of competitor bank pages to benchmark offers across the following parameters: value of offer, ease of fulfilment, quality of content or terms of usage
How can offer management be made easy for both merchant partners and internal teams?
- Leverage a single offer management platform that can ingest different offer types via flat files, APIs as well as make it self-serve for merchant partners and internal teams.
- Capture all data related to the offer such as merchant name, discount, coupon code, offer URL, T&C, validity, Address, Lat & long, redemption mechanism, etc are stored in a single repository.
- Ensure the platform receives daily updates from all sources.
How can managing merchant partnerships become easier?
Ensure contracts for offers are sourced by a single team or vendor. Ensure any offer types from any offer source can be ingested, cleaned and maintained on a single offer management platform.
How can customer experience become more engaging?
Here are some of ways used to make the marketplace web app more engaging:
- Marketing real estate:
- Support for campaigns and spot promotions
- Targeted campaigns
- Spot promotions
- Personalized landing pages
- Recommendations based on previous transactions
- Recommendations based on context (deals expiring, deals near the user, etc.)
- Recommendations based on interactions
- Real-time learning
- User preferences updated with every interaction
- Recommendations fine-tuned basis updated preferences
- Simplified discovery and redemption journey
- 3-step redemption of offers
- multi-language search on merchant attributes, names, offer details
- Relevant and superior content
- Wide choice of offers
- Local language content and offers
Should the experience be personalized for both pre-login and post-login journeys?
- Yes, the experience can be personalized in both pre-login and post-login journeys
- Pre login journey: The customer accesses the Web App through a generic URL. The identity of a customer is based on the device id. Each device id is considered as one unique customer. Day 0 personalization is based on popularity, trends, location etc. As customers start interacting with the offers - like, dislike, click through, views, redemption etc. these interactions should be used to improve the recommendations
- Post login journey: Customer identity is known, and APIs are used to return the most relevant offers based on the user’s context (such as location or time), and their preferences derived from historical transactions. In case a vendor is providing these capabilities then only masked customer identity should be sent to vendor’s API.
How should the handshake happen from bank’s digital assets?
- There are 3 potential options for Handshake:
- Option 1: (Recommended): Access PWA from within the bank’s Mobile app. In this approach, an authentication token can be passed to the PWA. The PWA can then be rendered within the mobile app in the in-app browser for seamless experience.
- Option 2: Provide access to customers through a Personalized URL (unique for each customer): Personalized URLs should be generated for each customer. These URLs can then be shared with customers through any of bank’s engagement channels such as SMS, WhatsApp, e-mail etc. Bank customers can directly access the Marketplace Web App through these URLs.
- Option 3: Provide access to customers through a Generic URL (common for all customers): A generic URL such as https://smartprivilegesoffers.com/ can be used to host the PWA. These URLs can then be shared with customers through any of bank’s engagement channels such as SMS, WhatsApp, e-mail etc. Bank customers can directly access the Marketplace Web App through these URLs.
How should a visit be attributed to a subsequent spend by the customer?
- An offer visit can be linked to a subsequent spend by the same customer in the following two scenarios:
- Online affiliate offers via marketplace: If a customer views and avails an offer on the web app, the transaction in the same session is tracked. If the transaction occurs in a separate session on the merchant's website, link it using date and time.
- Offline in-store offers: For in-store transactions, link the spend to the visit by mapping the date and time. This method may not capture all linkages but serves as a proxy for attributing spends to visits.
How should impact be measured?
- Setup reports to track the following:
- Customer Engagement metrics
- Spend metrics
- Campaign performance metrics
- App Analytics metrics
- Ensure all reports can be configured for an A/B testing scenario to measure the impact
How can new merchant co-fund partnerships be identified?
Identify merchants where customers have affinity to transact but don’t have offers yet. Estimate the potential of a merchant offer across customer coverage, transactions, and spends. Based on this information, decide which merchants to co-fund offers with.
Where should the offers be hosted?
- Offers should ideally be hosed on a Progressive Web App (PWA) for faster time to market and reducing the bandwidth needed from internal tech teams.
- This PWA if provided by a vendor should be white labelled for the bank, following its branding guidelines.
- Bank customers can access and interact with the offers through this PWA.
- PWA should use bank’s web domain so that customers don’t have to go to an external URL
- PWA can be hosted on any cloud infrastructure (AWS, Azure or GCP). Cloud instance could be owned by either Bank or Vendor
What offers data is to be captured? How will this be maintained?
All data related to the offer such as merchant name, discount, coupon code, offer URL, T&C, validity, Address, Lat & long, redemption mechanism, etc can be stored in a central database. This database should receive daily updates through APIs from affiliate sources and direct CSV file loads for merchants with direct partnerships.
What data is expected to be used by the bank? At what frequency?
The bank should ideally use the following data to build AI models to personalize the experience:
- Customer transaction data, once for a 12-month period, and daily increments (or as agreed).
- Customer demographic data, once, and weekly increments (or as agreed).
- Customer product holding data, once, and weekly increments (or as agreed).
- Incumbent offers data and merchant details, weekly or as offers become available (if applicable).
- Customer targeting or campaign data, weekly increments (or as agreed).
- If the bank leverages a vendor to build the AI models, then only masked data should be shared.
What customer data is needed to personalize the experience?
Ensure the experience is personalized based on:
- Context – location, day, date, time
- Clickstream data – views, likes, redemptions
- Historical transaction data
How can data hygiene be maintained on a regular basis?
Offer data hygiene can be maintained in different ways depending on the source of offer:
- All affiliates are connected via API. Ensure API pipelines are setup, run daily. Ensure there are automated checks and conditions set up while processing API to weed out any duplicate or inactive offers.
- For all direct partnerships, ensure there is a manual CSV load process involving maker checker and that the same file gets processed through the set checks and conditions before storing it in a central repository.
- Once the offer is ingested, ensure it is validated and enriched with additional attributes about the offer and merchant (e.g. image quality, data correctness, category tags).
- This process ensures high-quality data, and ability to promptly change/remove any offers from all front ends within 4-6 hours in case of issues or discrepancies.
Will customer PII be required?
No PII is required at any time.
Get Started Today
Have more questions? We’re here to help. Contact us to learn more, schedule a demo, or discuss custom solutions for your business.