Right Offer, Right Moment: Personalisation and Timing

Two banks can run the same offer at the same merchant and get very different results. The difference is rarely the offer itself. It is who sees it and when.

Personalisation decides whether an offer is relevant. Timing decides whether it is useful. Get both right and an offer feels like a helpful tip from a bank that knows you. Get them wrong and it is one more notification to dismiss.

Why generic catalogues underperform

Many programmes start with a catalogue: a long list of offers shown to every cardholder in the same order. It is simple to build, but it has predictable problems:

• most customers see offers for places they will never visit;
• the best offers for each customer are buried;
• customers learn that the offers section is not about them and stop checking.

Personalisation fixes the first two problems. Timing fixes the third, by giving customers a reason to come back at the moments when offers matter.

The signals that matter

Banks already hold rich, consented signals about how customers use their cards. The most useful for offers are:

• Category affinity: where a customer spends regularly, such as dining, groceries, fitness or travel.
• Location patterns: the neighbourhoods where they live, work and spend weekends.
• Frequency and rhythm: weekly routines, weekend habits and salary-cycle patterns.
• Life stage signals: a new card, a first travel transaction or a change in spending that suggests a move or a new family member.
• Engagement history: which offers they have viewed, saved and redeemed, and which they ignored.

None of this needs to be intrusive. The aim is not to know everything about a customer, but to avoid showing obviously irrelevant offers.

Personalisation in practice

Rank, don't restrict

The simplest improvement is to reorder, not hide. Every customer can still browse the full catalogue, but the first offers they see are the ones most likely to matter to them.

Mix relevance with discovery

If every offer matches past behaviour, customers only see more of what they already do. Good ranking keeps some room for discovery: a new restaurant near the office, or a wellness offer for someone who has started visiting a gym.

Segment for strategy

Personalisation also serves the bank's goals. A customer whose card is rarely used for groceries is a better target for a grocery offer than someone who already uses it every week, because the potential for incremental spend is higher.

Timing: the moments that convert

Even a perfectly relevant offer can fail if it arrives at the wrong time. Some moments work particularly well.

Before the routine

Show grocery offers before the usual weekly shop. Show dining offers on Thursday afternoon, before weekend plans are made. The offer arrives while the decision is still open.

Near the place

Location-based alerts, sent when a customer who has opted in is close to a participating merchant, catch the moment of greatest relevance. They must be rare, respectful and easy to turn off.

Around salary day

In many markets, including much of the GCC, spending rises after salaries are paid. Offers timed to the days after payday meet customers when they are ready to spend, and help the bank's card win those purchases.

Seasons and occasions

Ramadan, Eid, national days, school terms and summer travel all change what people buy and when. A campaign calendar built around these moments keeps offers timely without constant manual effort.

Guardrails that protect trust

Personalisation relies on trust. To keep it:

• Be transparent. Explain simply why customers see certain offers, such as "because you often dine in Seef".
• Respect consent. Location alerts and marketing messages need clear opt-in, and customers must be able to change their minds easily.
• Avoid sensitive inferences. Do not target offers based on health, financial difficulty or other sensitive signals.
• Limit frequency. Cap how many notifications any customer receives in a week.
• Keep data inside the bank. Merchants should see aggregated results, not individual cardholder data.

Where AI helps

Machine learning and AI assistants can make personalisation practical at scale. They can score which offers suit which customers, suggest audiences for a new campaign, and recommend the best day and time to send a message. In cardoff.ai, Hyduri drafts campaigns and suggests audiences in plain language, and a bank employee reviews and approves every campaign before it goes live.

The role of AI is to make good judgement faster, not to replace it.

Measuring the effect

To know whether personalisation and timing are working, compare:

• redemption rates of ranked versus unranked offer lists;
• conversion from notification to redemption at different times and days;
• incremental spend against a control group that saw a generic catalogue or no campaign;
• opt-out and unsubscribe rates, which signal when timing becomes intrusive.

The takeaway

Relevance and timing turn offers from a catalogue into a service. Use the signals you already have, rank rather than restrict, time offers to routines, places and salary cycles, and protect trust with clear guardrails. The result is fewer, better messages that customers actually act on.