AIDMA model
Review Attention, Interest, Desire, Memory, and Action to identify missing information and obstacles before purchase.
Use AIDMA as a lens for reviewing advertising and purchase communication. For each stage, consider the customer’s question, the touchpoint, and the obstacle. Unlike AIDA, it includes Memory. The stages are prompts for discussion, not proof that every customer follows the same path.
A target audience, purchase goal, communication touchpoints, and observations with clear event definitions.
Communication gaps, proposed changes, and a plan to observe the purchase journey.
How to use it
- Define the audience, purchase action, and observation period.
- Review what supports Attention, Interest, Desire, Memory, and Action.
- Compare the proposed path with actual questions, revisits, and skipped stages.
- Choose changes and track behavior with consistent definitions.
Keep its limits in mind
Page views, saves, and reservations may refer to different people and repeated events. Do not calculate a continuous conversion funnel without a consistent population and observation window.
Checking communication for a breakfast pack
Fictional example: Qinghe records 1,000 promotional-page views, 80 saves, and 20 reservations over one week. These event counts are not a matched cohort. Some interviewed customers are unsure whether pickup will fit their morning schedule.
On a small screen, scroll across the diagram to read it.
Interpretation: The counts do not prove a 2% conversion rate for one group or identify a psychological stage. The interview concern points to a pickup-information gap.
Proposed change: Clarify delivery and pickup arrangements and reduce reservation friction. No revised-page result is reported, and no time saving is measured.
Next step: Track views and reservations within a defined population and time window, and check whether customers understand pickup arrangements.
Three ways to misuse the five prompts
Use AIDMA to inspect purchase communication, not to diagnose a person’s mind. The practical question is whether someone can notice an offer, understand its relevance, remember useful information and take a clear next action. A person may return, skip a touchpoint or decide against buying for a sensible reason. A complete set of five boxes is a review aid, not evidence of a universal sequence.
Do not use it as a continuous conversion funnel when the counts refer to different populations or repeated events. Do not label a save as proven desire or a reservation as a completed purchase. Do not turn the Memory prompt into permission to send reminders. The permitted channel, purpose, consent where required and withdrawal mechanism must be established separately; a marketing worksheet supplies none of them.
Also separate information gaps from defects in the offer itself. If pickup times do not suit the intended customer, clearer copy may reveal the mismatch rather than remove it. That is useful learning. Avoid promising preparation-time savings without measurements or inventing urgency to force action. The offer, price, terms and practical arrangements should remain understandable even when the customer chooses not to proceed.
Choose between message review and observed experience
AIDMA helps generate questions about communication. Use a different structure when the unresolved issue is an observed service sequence, an undefined audience or a set of comparable product measures.
| Method | Useful input | Output to expect |
|---|---|---|
| AIDMA | An audience, offer, touchpoints and open customer questions | Gaps in purchase communication and proposed checks |
| Customer journey map | A defined user situation and observed steps | Experience, obstacles and evidence gaps across touchpoints |
| STP | Research about segments and a bounded market | A reasoned target and positioning hypothesis |
| KPI tree | Consistently defined measures within compatible windows | A numerical decomposition, not a psychological explanation |
Worked continuation: keep counts and stories separate
In the fictional breakfast-pack case above, the week contains 1,000 promotional-page views, 80 saves and 20 reservations. They are unmatched event counts. One visitor could create several views or saves, and a reservation could arrive through another route. Therefore 20/1,000 is a ratio of these event counts, not an established 2% conversion rate for one group of people. No completed-purchase count or measured time saving is supplied.
For the qualitative part of this exercise, consider three additional invented interview fragments. They are illustrative stories, not a sample linked to those weekly counts. P1 heard about the pack from a colleague and went straight to the reservation link without saving the page. P2 saved the offer twice, then declined after seeing the pickup window. P3 returned to compare contents but still could not tell whether the pack suited two people. These stories demonstrate possible questions; they do not estimate how frequently each path occurs.
Map each fragment to the communication needed, without assigning the person to a fixed stage. P1 challenges the assumption that Memory must appear as a save. P2 shows that repeated saves do not prove purchase intent. P3 identifies a content question that can be checked directly. Keep the existing concern about morning pickup visible rather than replacing it with a generic instruction to make the page more persuasive.
| Material | Question raised | Proposed response and check |
|---|---|---|
| P1: direct reservation after a colleague’s suggestion | Can the action page stand on its own? | Put contents, terms and pickup information at that destination; check understanding |
| P2: two saves, then declines | Does the service fit the available morning window? | Clarify the actual window; investigate fit before proposing a service change |
| P3: returns to compare contents | Is the two-person offer understandable? | Show accurate contents and ask the reader to explain what is included |
| Weekly 1,000 / 80 / 20 events | Are population and event definitions compatible? | Define measurement before calculating a customer conversion rate |
Review comprehension before claiming a lift
The conclusion is that pickup information and pack contents deserve checking, while the available numbers cannot locate a psychological bottleneck. Request a limited communication review rather than claiming that “Memory is weak.” Draft clear pickup wording, verify it with operations and ask intended readers to explain the arrangements back in their own words. Record confusion, correct understanding and people for whom the offer simply does not fit.
If a change is tested, define the action being measured: reservation submitted, reservation accepted, pack collected or paid purchase. State whether counts represent unique eligible people, sessions or events, the observation window and how duplicates are handled. Collect only what is needed under the applicable privacy arrangements; do not introduce personal identifiers into advertising requests to make the analysis convenient.
At review, preserve the old and revised wording, what actually changed, the population observed and any concurrent service changes. Better comprehension does not by itself prove more purchases, and more reservations do not prove saved preparation time. Record reminder settings and actual permission separately. A decision to improve service hours requires its own capacity and approval discussion; the message review has not granted it.
Check your work
- Each stage has a relevant touchpoint or question.
- Customer concerns are distinct from proposed responses.
- Reminder consent and delivery conditions are clear.
- Event counts are not silently treated as unique people or a single cohort.
Put the framework to work
Copy the outline and add your question, evidence, and open questions.