Meta Ads Interview Questions for Experienced Marketers

Browser tracking loses a large share of events to ad blockers and privacy limits, which starves optimisation. The Conversions API sends events server-side to recover them. You run both, with a shared event_id so Meta deduplicates and each conversion counts once.

A score for how well your events’ customer data lets Meta match events to accounts. Improve it by sending more hashed parameters, email, phone, name, IP, user agent, which lifts attribution and optimisation, especially via the Conversions API.

App Tracking Transparency let users opt out, cutting signal sharply. Meta responded with Aggregated Event Measurement, which models opted-out conversions. The manual eight-event ranking step was automated away in 2025, and numbers are now partly modelled, so treat them as directional.

The window crediting a conversion to an ad, for example 7-day click and 1-day view, the current default. A longer window credits more conversions to ads, so you always compare like-for-like and sanity-check against real sales.

When the same event fires from both browser and server, Meta uses a shared event_id (and event name) to recognise them as one and drop the duplicate. Without it you double-count conversions and mislead optimisation.

For Special Ad Category campaigns where it is required, brand-new accounts with no data, hyper-local or niche B2B, very small budgets, or when you need a hard boundary on who is reached. Otherwise Advantage+ usually wins with data and good creative.

You feed Meta a source that includes purchase value, so it weights toward high-value customers rather than any customer. Use them for e-commerce to find people likely to spend more, not just convert.

Because Meta’s AI now finds buyers better than manual interest stacks, and detailed-targeting options were cut. Feeding strong signals and letting Advantage+ explore a broad pool typically beats narrow, hand-built audiences on cost per result.

Segment by warmth, video viewers and engagers, then site visitors, then cart or form starters, then past buyers, and match the message to each stage. Warm audiences convert several times cheaper, but you must keep the top of funnel fed.

Meta removed detailed-targeting exclusions earlier and cut a large set of interest options in January 2026. Interests now act as a suggestion the AI can exceed, not a hard filter, pushing strategy toward broad plus strong creative.

Because each ad set needs about 50 events a week to exit the learning phase. Many small ad sets fragment budget and data so none stabilise. Fewer, well-fed ad sets give cleaner signals and cheaper, steadier delivery.

Formerly Advantage+ Shopping, it is the fully automated campaign type for sales, app and lead objectives, handing most decisions to the AI. Use it for e-commerce and scaled prospecting, feeding it strong creative and a clean catalog or pixel.

Use the built-in A/B test tool so audiences split without overlap, change one variable at a time, and turn off Advantage+ Campaign Budget so each variant gets equal spend. Give each enough volume, near the 50-event guide, before judging.

Top of funnel with Awareness or Traffic to reach cold audiences, middle with engagement and retargeting to nurture, and bottom with Leads or Sales to convert. Each stage builds the audience for the next, so you keep feeding the top.

A consistent convention encoding what you filter by, for example Objective_Offer_Date at campaign level and Audience or Format_Version below. It keeps reporting scannable and makes scaling and pausing losers painless.

Highest Volume (most results, no cost control, the default), Highest Value (maximise purchase value), Cost Per Result Goal (target average cost, formerly Cost Cap), ROAS Goal (target return, formerly Minimum ROAS), and Bid Cap (a hard per-auction ceiling, strictest).

Cost Per Result Goal targets an average cost and is more forgiving, good once you know your CPA. Bid Cap sets a hard ceiling on every bid for strict control, but needs solid data and can choke delivery if set too low.

Vertically, raise the budget about 20% every few days to preserve learning, or horizontally, duplicate the winner into new audiences, geographies or creatives. Big sudden budget jumps reset learning and can wreck performance.

Splitting spend across many ad sets means none reach the roughly 50 weekly events needed to exit learning, so all stay expensive and volatile. Consolidating into fewer, better-funded ad sets is usually more efficient.

Often a cost or ROAS target set beyond what the auction allows, so Meta cannot deliver. Other causes include a tiny audience, an exhausted schedule, a spending limit hit, or rejected ads. Base caps on real data, not wishes.

Fix it structurally, not with patience: raise the budget, broaden the audience, consolidate ad sets, or optimise for a more frequent event using the event ladder. It means Meta cannot get about 50 events a week as configured.

The problem is almost always after the click: a slow or broken landing page, a weak offer, or an ad promise the page does not keep. Check message match and page speed before touching the creative or audience.

The signature is rising frequency, falling CTR from its peak, and rising CPA with nothing else changed. The audience has seen the ad too often. The fix is fresh creative on a testing rhythm, not a new audience.

Test big levers in order, hook, offer, format, then details, one variable at a time, with enough budget per variant to gather real results. Keep a steady cadence: mostly iterate winners, some new hooks, a little experimental.

Work systematically: check for recent edits that reset learning, ad rejections or disapprovals, rising frequency and fatigue, a broken pixel or landing page, seasonality or auction competition, and audience overlap. Isolate one variable, confirm against real sales, then act.