framework · strategy
AIDA framework for product management interviews
Best for: The messaging and communications layer inside a GTM or product launch answer. Not a complete GTM framework. Use it to structure what you say to whom at each awareness stage, then go broader for pricing, distribution, and success metrics.
AIDA is a 128-year-old advertising model, created by Elias St. Elmo Lewis in 1898 for single-buyer, linear purchase journeys. That context matters in a PM interview. It is a communications framework, not a business strategy framework. It answers “what do we say to whom at each stage of awareness?” It does not answer “how do we price this, how do we distribute it, or how do we measure launch success?” Candidates who present AIDA as a complete GTM answer send a weak signal. Candidates who use it precisely as the messaging layer, then add pricing, channel strategy, and stage-level metrics around it, send a strong one.
The four stages with PM-specific definitions
Attention: Make the specific target user aware the product or feature exists. The interview question is not “which channel has the most reach?” but “where does this pain actually surface for this person, and can we appear there?” An in-product tooltip at the moment a user hits a workflow failure is a higher-leverage attention mechanism than a cold email blast, because it meets them at the point of relevance. Name the segment, name the channel, and say why.
Interest: Demonstrate that the product is relevant to their specific problem. Generic is the enemy here. “We’ll write blog posts and run demos” fails because it could apply to any product in any category. The question is: what does this user currently fail at, and how do we show them a concrete before/after tied to that failure? Interest is earned by specificity, not volume.
Desire: Create preference over the next-best alternative, which usually means addressing the real switching cost or trust objection. In 2026, desire is rarely built through social proof and testimonials alone. For AI products especially, desire is earned by proving the product is viable (it solves a recurring, real problem someone will pay to fix) and lovable (it meets users in their existing workflow without adding cognitive load). A demo that impresses but requires a behavior change creates weak desire. The lovable, not just usable framing matters most at this stage.
Action: Reduce friction to the first meaningful action. This is not sign-up. It is the action that delivers the core value: the first successful output, the first completed workflow, the first moment the product earns trust. Instrument that specific moment and optimize the path to it. If you name “free trial” as the action, the interviewer hears that you have not thought through what the product actually needs users to do.
A fifth stage: Retention
At growth-focused companies (Meta, DoorDash, Lyft), interviewers will expect you to address what happens post-action. Add Retention explicitly: what is the habit loop that brings users back, and which metric tells you the loop is working? For consumer products, this is usually a D7 or D30 retention cohort. For B2B, it is weekly active users against the seat count. Do not leave this implied.
Worked example: AI writing assistant launch for legal teams
Prompt: “How would you launch an AI contract drafting feature for existing customers of a legal workflow platform?”
Start by flagging scope. “I’ll use AIDA to structure the messaging layer, but I want to address pricing, channel strategy, and metrics separately before we close.”
Attention: “The target is in-house legal counsels at mid-market companies who already use the platform for document storage. The highest-leverage attention channel is not email: it’s an inline prompt inside the document creation flow, surfaced the first time a user opens a blank contract template after the feature ships. That’s where the pain is. We do not need them to remember a marketing email they received two weeks ago.”
Interest: “We show them one before/after: a standard NDA that previously took 45 minutes to draft from a template now takes four minutes with AI-generated first draft plus inline clause suggestions. We tie this to a specific job they already do weekly, not a feature list.”
Desire: “The real objection is trust, not awareness. Legal professionals cannot risk a hallucinated indemnification clause. Desire comes from showing audit trails, accuracy rates against a benchmark clause library, and a one-click ‘send for human review’ escape hatch. We instrument early accuracy data and surface it inside the product, not in a marketing email.”
Action: “The first action is completing a full draft with the AI, not activating the feature. We instrument the first-draft-to-accepted event. The path from in-product prompt to first accepted draft has at most two steps. If users are dropping between feature discovery and first draft completion, that path is too long.”
Retention: “The habit loop is weekly: users who draft at least one AI-assisted contract per week in the first 30 days have an 8x higher 90-day retention rate (hypothetically, based on comparable workflow tools). Our leading indicator is weekly AI drafts per seat, not logins.”
Then close on metrics per stage: impressions on the in-product prompt, feature activation rate, time-to-first-draft, draft acceptance rate, and D30 weekly active users against activated users.
Strong and weak answers
strong
"Before I go to AIDA, let me flag its scope: it covers the messaging layer, not the full GTM. I'll use it for that, then cover pricing and channel rationale separately. For Attention, my target is [specific segment]. The highest-leverage channel is not the one with the most reach, it's the one where the problem surfaces. That's [specific channel and why]. For Interest, I'm showing them [specific before/after tied to their actual job], not a feature list. For Desire, the real objection is [switching cost or trust issue], and I address it by [specific mechanism, not 'testimonials']. For Action, the target event is not sign-up, it's [the first moment of core value delivery]. I'd instrument that specifically and optimize the path to it. Post-action, I'd add Retention: the habit loop is [specific], and the leading indicator is [specific metric]. The stage-level metrics I'd track: [top-of-funnel impressions, activation rate, time-to-first-value, NPS driver score by cohort]."
weak
"I'd use the AIDA framework. First, I'd create awareness through social media and PR. Then I'd build interest with blog posts and demos. Then I'd create desire with testimonials and case studies. Finally, I'd drive action with a free trial and a clear CTA." This fails on four counts: there is no named segment, no channel rationale, no metrics at any stage, and no acknowledgment that AIDA is only the messaging layer. The interviewer cannot evaluate whether the candidate knows what success looks like or whether they can reason through a real GTM.
When AIDA is the wrong tool
AIDA was designed for single-buyer, linear journeys. It fits poorly in three situations. First, complex B2B sales with multiple stakeholders: a CFO, an IT buyer, and an end user all need different messages, and they move non-linearly. AIDA’s single funnel cannot accommodate multi-stakeholder dynamics. Second, platform and API products where adoption happens through developers or partners, not end users. Third, viral consumer products where discovery happens peer-to-peer inside an existing product, not through a staged awareness campaign. In these cases, reach for the 4Ps or working backwards to structure the broader GTM, and use AIDA only if there is a clear awareness campaign component to address.
The 2026 reframe
AIDA’s weakest stage in 2026 is Attention. For many AI features, attention is earned passively through ambient integration: a copilot surfaced inline at the moment of need, an agent that surfaces a suggestion without being invoked. The traditional campaign-driven attention model may not apply. If your GTM requires users to remember a product exists and choose to go find it, you have already lost a significant share of potential adopters.
Desire has also shifted. When feasibility is near-free, desire is no longer built by demonstrating what the product can do. It is built by proving it solves a recurring, paying-customer-level problem (viability) and that it fits the existing workflow without creating new cognitive overhead (lovability). A product that impresses in a demo but requires behavior change creates weak desire. The bar is: would this user reach for this tool without being reminded it exists? If the answer is no, the desire stage has not been won. See proving viability for how to make this argument concrete in an interview answer.