Real questions from interviews at Ogilvy, Accenture, WPP Media, Cognizant, Dentsu India & Omnicom Media. After sitting through back-to-back Performance Marketing interviews at six leading agencies and consultancies, I walked away with one clear insight: top employers are not hiring ad managers — they are hiring marketers who can think in systems.
The questions go deep. Bidding logic, attribution philosophy, pixel architecture, funnel strategy — nothing is off limits. To help you prepare, I have documented every question asked across these interviews, paired with answers that reflect what actually impressed interviewers.
Whether you are entering your first agency role or making a senior move, this guide covers everything you need.
Section 1: Google Ads, Meta Ads & Performance Marketing Fundamentals
Q1. What are the latest features introduced in Google Ads & Meta Ads platforms?
Google Ads — Key Updates:
AI is now central to how Google Ads operates. The headline features shaping campaigns today include Performance Max (PMax), which distributes ads across every Google surface — Search, Display, YouTube, Gmail, Maps, and Discover — from one unified campaign driven entirely by machine learning. Demand Gen campaigns replaced the older Discovery format, prioritising visual storytelling across YouTube Shorts and Google’s content feeds. Generative AI now assists with asset creation (headlines, descriptions, images), and Google has positioned Broad Match paired with Smart Bidding as the recommended default setup for most accounts.
Meta Ads — Key Updates:
Meta’s response to the iOS 14 privacy shift has defined its product direction. The Advantage+ suite is the centrepiece — Advantage+ Shopping Campaigns automate the full e-commerce funnel, Advantage+ Audience lets Meta’s AI identify converters without manual targeting constraints, and Advantage+ Creative dynamically adapts visuals per placement. Reels has emerged as the highest-performing placement on CPM efficiency, and Threads Ad placements are now live. Conversion API (CAPI) is no longer optional — it is essential for accurate server-side tracking.
Interview angle: Anchor your answer to the AI and privacy-driven shifts on both platforms. Interviewers want to see you understand direction, not just features.
Q2. What is an automated bidding strategy? Can we fully rely on Smart Bidding?
Automated bidding uses machine learning to determine the ideal bid for every individual auction in real time. Rather than setting a static CPC or CPM, the system reads dozens of contextual signals — device type, location, time, browsing history, audience overlap — and calculates a bid aligned to your goal.
Google Smart Bidding options include Target CPA, Target ROAS, Maximise Conversions, Maximise Conversion Value, and Enhanced CPC.
Can you fully rely on it? Not without conditions. Smart Bidding produces strong results when:
- The campaign generates at least 30–50 conversions per month
- Conversion tracking is precise and reflects actual business outcomes
- The account structure is consolidated (fragmented micro-campaigns starve the algorithm)
- The learning phase (2–4 weeks) is protected from major changes
It underperforms when accounts are new, conversion volume is low, tracking is misconfigured, or when lead quality matters more than raw lead volume. A hybrid model — automated bidding for established campaigns, manual CPC or Enhanced CPC for testing and low-data scenarios — remains the most reliable approach.
Q3. What is an effective audience segmentation strategy in Performance Marketing?
Strong segmentation goes beyond demographics. It layers intent signals, behavioural data, and customer lifecycle stage to match the right message to the right person at the right moment.
A practical framework:
Funnel-stage segmentation separates cold prospects (TOFU), engaged non-converters (MOFU), and high-intent users (BOFU) into distinct campaign groups with tailored messaging and bids.
Behavioural segmentation uses on-site event data — pages visited, scroll depth, CTA interactions — to identify where users are in the decision process.
Demographic overlays (age, gender, location, income) refine broad audiences rather than anchor them.
First-party data segmentation — CRM lists bucketed by purchase history, LTV tier, or lead score — produces the highest signal quality of any audience type, as the data originates from your own customer relationships rather than platform inference.
The primary purpose of segmentation is message differentiation, not just targeting efficiency. A cart abandoner and a first-time visitor require completely different creative approaches.
Q4. What are custom conversions, and how do you track and analyse them across multiple funnels?
Custom conversions are marketer-defined tracking events built to capture actions that matter to a specific business — a thank-you page visit, a high-value purchase threshold, a mid-funnel video completion — rather than relying solely on platform defaults.
Tracking across multiple funnels works best with this structure:
Google Tag Manager handles the trigger logic — individual tags fire for each funnel milestone (form view, form start, submit, qualified lead). GA4 custom events map to each stage and feed back into Google Ads as conversion actions. UTM parameters on every paid URL ensure GA4 correctly attributes each conversion to its originating campaign and channel.
On Meta, Events Manager allows custom event creation beyond standard events, with testing tools to verify that each fires correctly.
Analysing top performers: In Google Ads, segment the Conversions column by conversion action to compare CPA and volume across funnel stages. In GA4, the Funnel Exploration report surfaces drop-off points. Cross-referencing ROAS and CRM pipeline data identifies which campaigns drive volume versus actual revenue.
Q5. What is Attribution Modelling?
Attribution modelling defines how credit for a conversion is distributed across all the touchpoints a user engaged with before completing an action. Most customers do not convert on first contact — they move through multiple channels and sessions before deciding.
Common models and their logic:
| Model | Credit Allocation | Best Application |
|---|---|---|
| Last Click | 100% to the first touchpoint | BOFU conversion analysis |
| First Click | 100% to first touchpoint | Awareness channel valuation |
| Linear | Equal split across all touchpoints | Balanced multi-channel view |
| Time Decay | More weight to recent touchpoints | Short sales cycles |
| Position-Based | 40% first, 40% last, 20% middle | Acquisition-focused brands |
| Data-Driven | ML-based on actual conversion patterns | High-volume, data-rich accounts |
GA4 defaults to Data-Driven Attribution. The model choice directly affects budget decisions — last-click attribution consistently undervalues awareness channels that initiate the conversion journey.
Q6. What is campaign optimisation? What are the pros and cons of automation in paid campaigns?
Campaign optimisation is the continuous process of improving performance against business KPIs — adjusting bids, budgets, creatives, targeting, and landing pages based on data signals in an ongoing cycle.
Advantages of automation: Real-time bid adjustments at a scale no human team can replicate. Faster adaptation to seasonal patterns and audience behaviour changes. Reduced manual workload, freeing teams for strategy and creativity. Strong performance outcomes when conversion data is abundant.
Disadvantages of automation: Limited transparency into algorithmic decision-making. A 2–4 week learning period with performance volatility. Weak results when the monthly conversion volume falls below 30. Risk of optimising for conversion quantity over quality if downstream lead data is not fed back into the platform. Over-reliance on automation can mask creative fatigue and structural inefficiencies.
The right posture: use automation as a tool that amplifies good strategy — not as a substitute for it.
Section 2: Campaign Structure & Strategy
Q7. Can you run multiple campaign objectives for the same audience?
Yes, but it requires careful management. Running two campaigns with different objectives against the same audience creates internal auction competition — both campaigns bid against each other for the same users, inflating CPMs and wasting budget.
The fix: apply audience exclusions at the ad set level (Meta) or use negative audience lists (Google) so each campaign owns a distinct segment. Use Meta’s Audience Overlap tool before launch to check shared reach between any two audiences. If overlap exceeds 30%, consolidate campaigns or enforce hard exclusion rules.
Q8. How many bidding strategy types exist in Google Ads and Meta Ads?
Google Ads (9 strategies): Manual CPC, Enhanced CPC, Maximise Clicks, Target Impression Share, Target CPA, Target ROAS, Maximise Conversions, Maximise Conversion Value, CPM/vCPM (Display/YouTube).
Meta Ads (4 strategies): Lowest Cost (automatic), Cost Cap (average ceiling per result), Bid Cap (hard per-auction maximum), Minimum ROAS (for catalogue/shopping objectives).
Q9. What happens when two campaigns target identical audiences with the same objective? How do you fix overlap?
Internal auction competition drives up costs — you are essentially bidding against yourself. CPMs rise, budget efficiency drops, attribution data gets double-counted, and overall CPA increases without any change in actual market demand.
Fixes include: applying mutual audience exclusions between campaigns, consolidating overlapping campaigns into one (Google and Meta both recommend fewer, larger campaigns), redefining campaign scope so each owns a unique funnel stage, and using Meta’s Audience Overlap checker or Google’s Audience Manager to audit before and after structural changes.
Q10. Explain a complete marketing funnel strategy.
TOFU — Awareness: Objective is reach and recall. Campaigns: YouTube/Reels video, Google Display, Meta Reach. Audience: broad interest + 3–5% lookalikes. Creative: problem-aware short video. KPIs: impressions, CPM, view rate.
MOFU — Consideration: Objective is education and intent capture. Campaigns: Google Search (non-brand), Meta Traffic/Lead Gen, YouTube in-stream. Audience: TOFU engagers, website visitors (30–60 days). Creative: value proposition, testimonials, demos. KPIs: CTR, CPL, landing page conversion rate.
BOFU — Conversion: Objective is action — purchase, sign-up, booked call. Campaigns: Google Search (brand + competitor), Meta Conversion, retargeting. Audience: cart abandoners, high-intent page visitors (7–14 days). Creative: urgency, offer, objection-handling. KPIs: CPA, ROAS, conversion rate.
Retention: Objective is repeat purchase and LTV growth. Campaigns: dynamic retargeting to past buyers, Google Customer Match, email-reinforced paid ads. KPIs: repeat purchase rate, upsell revenue.
Q11. If broad targeting performs well, why do we still need custom audience segmentation?
Broad targeting — especially Meta’s Advantage+ Audience and Google’s broad match — has genuinely improved. In data-rich accounts, algorithm-driven targeting often outperforms manually defined interest audiences.
But segmentation remains essential for four reasons:
Messaging differentiation: the algorithm can find the right person, but it cannot show them a different message based on where they are in the funnel. A cold prospect and a returning visitor need different creative — that distinction requires segmentation.
Bid differentiation: not every customer segment has equal value. Segmentation allows you to bid higher for high-LTV cohorts.
Exclusion management: Without segmentation, you cannot reliably exclude existing customers from acquisition campaigns or avoid re-targeting users who converted yesterday.
Insight generation: Broad campaign data produces averages. Segmented data reveals which cohorts drive the most value — information that is essential for scaling decisions.
Q12. Explain a lead generation funnel strategy.
Stage 1 — Attract (TOFU): Meta/YouTube video ads to broad interest or lookalike audiences. Goal: awareness. Creative: problem-framing content, no selling. KPIs: CPV, reach.
Stage 2 — Engage (MOFU): Retarget video viewers (25%+) and website visitors with a lead magnet offer — guide, webinar, free audit, or demo. Use Meta Lead Gen forms or a dedicated landing page. KPIs: CPL, form completion rate.
Stage 3 — Convert (BOFU): Retarget form openers who did not submit. Use Google Search for branded and competitor queries. Creative: case studies, urgency, and FAQ-handling. KPIs: qualified CPL, lead-to-booked-call rate.
Stage 4 — Nurture: Pass leads to CRM via direct integration. Layer email sequences with paid remarketing using CRM-based custom audiences. KPIs: lead-to-SQL rate, pipeline contribution, revenue per lead.
Section 3: Analytics, Tracking & Optimisation
Q13. The difference between manual and automated bidding. What is the trade-off?
| Factor | Manual Bidding | Automated Bidding |
|---|---|---|
| Control | Full | Algorithm-controlled |
| Transparency | High | Limited (black box) |
| Speed of adjustment | Slow | Instant, per-auction |
| Data requirement | Any volume | 30–50+ conversions/month |
| Learning period | None | 2–4 weeks |
| Best for | New accounts, testing, brand protection | Scaled, data-rich campaigns |
The core trade-off: manual offers control at the cost of scalability; automated offers performance optimisation at the cost of visibility. In practice, a portfolio approach — automated for mature campaigns, manual for early-stage or niche campaigns — produces the best outcomes.
Q14. What is frequency in Meta Ads?
Frequency is the average number of times each unique person in your audience has seen your ad.
Formula: Frequency = Total Impressions ÷ Total Reach
A frequency below 2 typically means insufficient brand exposure. The 3–7 range is generally where awareness converts to consideration. Above 8, ad fatigue sets in — CPMs rise, CTR falls, and users begin hiding your ads, which negatively impacts delivery quality and future costs. Monitor CPM and CTR trends together weekly; a simultaneous rise in CPM and fall in CTR is the clearest signal of fatigue.
Q15. What is the difference between frequency and reach?
Reach counts the total number of unique individuals who saw your ad at least once — it measures breadth.
Frequency measures how many times each of those individuals saw it on average — it measures depth.
Together: Reach × Frequency = Total Impressions.
For awareness campaigns, prioritise reach with low frequency (1–3). For retargeting, a higher frequency (5–8) is appropriate because the audience is already warm and needs repeated exposure to convert. Neither metric is meaningful in isolation.
Q16. Explain GA4 events, conversions, and attribution reports.
GA4 Events fall into four categories: automatically collected events (page_view, session_start), enhanced measurement events enabled in settings (scroll, outbound click, video engagement), recommended events following GA4’s naming schema (purchase, generate_lead, sign_up), and custom events created for business-specific actions via GTM or gtag.js.
Conversions in GA4 are simply events marked as conversions via Configure → Events. Any event can be a conversion — this feeds into Funnel Exploration reports and the linked Google Ads property.
Attribution Reports live under Advertising → Attribution → Model Comparison. This is where you compare how different attribution models (Last Click, First Click, Data-Driven, etc.) redistribute credit across channels — and see which channels are being undervalued by the model currently used in campaign reporting.
Section 4: Performance Metrics & Reporting
Q17. Difference between Search campaigns and Performance Max. Have you used PMax in a real project?
Search Campaigns run exclusively on Google Search. You control keywords, match types, ad copy, bids, and targeting. High intent — the user is actively searching. Full keyword-level performance transparency.
Performance Max runs across all Google inventory from a single campaign. AI controls placement, bidding, creative combinations, and channel allocation. You provide asset groups, audience signals, budget, and a conversion goal. Transparency is limited — Insights shows search themes, not individual queries.
In practice: PMax works best as a complement to Search, not a replacement. Run a dedicated Brand Search campaign alongside PMax to prevent it from cannibalising branded traffic. Add account-level negative keyword lists. Use your best customer list as the primary audience signal to accelerate the algorithm’s learning.
Q18. CTR vs. Conversion Rate — what is the difference?
CTR = (Clicks ÷ Impressions) × 100. Measures ad relevance and creative strength.
Conversion Rate = (Conversions ÷ Clicks) × 100. Measures landing page quality, offer strength, and audience intent fit.
Diagnostic approach:
- High CTR + Low CVR → compelling ad, but landing page or audience intent is misaligned
- Low CTR + High CVR → poor creative reach, strong buyer intent in those who do click
- Low CTR + Low CVR → fundamental audience or offer problem
- High CTR + High CVR → healthy full-funnel performance; scale it
Optimise CTR through ad creative. Optimise CVR through landing page, offer, and audience quality. Neither metric tells the full story alone.
Section 5: Google Ads & Meta Ads Implementation
Q19. Explain Google Ads account structure and campaign hierarchy.
Google Ads operates on four levels:
Account level: Billing, currency, time zone, linked GA4 property, conversion tracking, shared libraries (negative keyword lists, audience lists, shared budgets).
Campaign level: Objective, campaign type (Search, Display, Shopping, Video, PMax), budget, bidding strategy, geographic and language targeting.
Ad Group level: Keyword themes (Search) or audience/topic targeting (Display). Each ad group should contain a tightly themed keyword cluster. Bids can be adjusted at this level.
Ad level: Responsive Search Ads with up to 15 headlines and 4 descriptions. Google tests combinations automatically. Pin critical headlines (brand name, primary CTA) to fixed positions; leave others unpinned for testing. Aim for an Ad Strength rating of Excellent.
Q20. Explain the campaign lifecycle and optimisation workflow.
Phase 1 — Launch: Define KPIs, verify conversion tracking, build structure, set conservative initial bids.
Phase 2 — Learning (Weeks 1–2): Avoid major structural changes. Monitor tracking accuracy, spend pacing, and placement relevance.
Phase 3 — Optimisation (Weeks 2–6): Review search term reports weekly (add negatives), A/B test ad copy, analyse device/location/time performance, pause underperformers.
Phase 4 — Scaling (Month 2+): Increase budget in 15–20% increments once CPA/ROAS is stable. Expand keyword themes, audience segments, and creative variants.
Phase 5 — Review (Ongoing): Monthly attribution review, creative refresh every 4–6 weeks, quarterly full account audit covering structure, bidding strategy, and competitive positioning.
Q21. What is Performance Max campaign automation?
PMax automates bidding, placement selection, creative assembly, and channel allocation using Google’s AI. You provide asset groups (headlines, images, videos, logos), audience signals (not hard targeting — algorithmic hints), a budget, and a conversion goal. Everything else is determined by the platform.
Key limitations: brand keyword cannibalisation (fix with a separate brand Search campaign), limited search query transparency (Insights tab shows themes only), and no native campaign-level negative keywords — these must be submitted via Google Ads support or set at the account level.
PMax requires high-quality creative inputs — accounts that provide diverse image formats and original video assets see significantly stronger performance than those relying on auto-generated assets.
Q22. Audience exclusions, custom audiences, and lookalike audiences — what are the differences?
Audience Exclusion: Prevents specific segments from being served your ads. Critical for acquisition efficiency — exclude existing customers, recent converters, or internal employees from prospecting campaigns.
Custom Audience: Built from first-party data — CRM email/phone uploads, website visitors (pixel-based), video viewers, app users, or engagement-based segments. Highest signal quality because the data reflects real behaviour with your brand.
Lookalike Audience: Algorithm-generated audience resembling your custom audience source. Meta’s 1%–10% scale (1% = highest similarity, 10% = broadest reach). Best seed audiences: high-LTV purchasers → all purchasers → qualified leads. Use 1–2% for conversion campaigns, 4–5% for broader prospecting.
Q23. What is the difference between remarketing and retargeting?
Retargeting specifically refers to pixel-based paid advertising served to users who previously visited your website or app. It is executed through Google Display, Meta, or programmatic channels using browser cookies or SDK events.
Remarketing is a broader term covering all re-engagement strategies — email sequences, SMS, CRM-triggered outreach, direct mail, and paid retargeting ads. In Google Ads, “Remarketing” is also the official term for their pixel and customer list-based targeting product (RLSA — Remarketing Lists for Search Ads).
In agency conversations, the terms are used interchangeably. Knowing the distinction signals precision.
Q24. What is conversion tracking and what are its types?
Conversion tracking connects ad spend to business outcomes and powers every Smart Bidding algorithm. Without accurate conversion data, automated bidding optimises toward the wrong goals.
Types:
Website conversions — form submissions, purchases, button clicks, thank-you page visits. Tracked via Google Tag/gtag.js, Meta Pixel, or GTM.
Phone call conversions — calls directly from call extensions or Google forwarding numbers on-site.
App conversions — installs, in-app purchases, registrations. Tracked via Firebase (Google) or mobile measurement partners (AppsFlyer, Adjust).
Offline / imported conversions — CRM conversions (lead → qualified sale → closed deal) imported back into Google Ads using GCLID matching via the Offline Conversion Import API. Essential for B2B.
Store visit conversions — Google estimates physical visits from ad exposure using location history.
Lead form conversions — completions of native in-platform forms (Google Lead Form Extensions, Meta Instant Forms).
Q25. What are the main audience targeting types — Interest, Custom, Lookalike & Remarketing?
Interest Targeting: Platform-inferred audience categories based on content consumption and engagement history. Broad reach, lower intent. Suited to TOFU prospecting when first-party data is limited.
Custom Audience: First-party data segments — CRM lists, pixel-based website visitors, video viewers, app users. Highest precision. Used for retargeting, exclusions, and as seeds for lookalikes.
Lookalike Audience: Platform-generated audience that mirrors your custom audience. Bridges the gap between first-party precision and prospecting scale. Quality depends entirely on the source — always use your best-performing customer segment as the seed.
Remarketing Audience: Prior brand engagers — website visitors by URL, product viewers, cart abandoners, video watchers, page followers. The most conversion-ready audience type. Requires urgency-driven creative with specific offers or objection-handling.
Section 6: Meta Ads, Pixel & Tracking
Q26. Explain basic Meta Ads targeting.
Meta Ads targeting is set at the Ad Set level and operates across three audience types.
Core Audiences combine location (country, city, or radius), age, gender, language, and Detailed Targeting (interests, behaviours, demographic qualifiers). Note: Meta removed many sensitive interest categories from 2022 onwards.
Custom Audiences include Website Custom Audiences (pixel-based), customer list uploads, video engagement audiences (watched 25/50/75/95/100%), page and Instagram engagers, and lead form openers.
Lookalike Audiences range from 1% (tightest similarity) to 10% (broadest scale). Recommended approach: 1–2% for conversion campaigns, 3–5% for awareness and consideration.
Advantage+ Audience is Meta’s AI-driven default targeting option. You can provide optional audience suggestions as guardrails, but the algorithm identifies who actually converts. Works best when the pixel has recorded at least 1,000 monthly events and a clear conversion signal is set.
Q27. How have you used Meta Ads in your projects?
In a B2C e-commerce engagement, I structured Meta campaigns across three funnel stages:
At the awareness stage, I ran Reels and single-image video ads targeting a 2–3% lookalike built from the top 500 purchasers by order value, overlaid with broad lifestyle interests. This accounted for roughly 20% of total Meta spend.
At the consideration stage, I retargeted 30-day website visitors (excluding purchasers) and video viewers at 50%+ watch completion using carousel ads with dynamic UTM-tagged product links. Budget: approximately 30% of total spend.
At the conversion stage, dynamic product ads (DPAs) ran against cart abandoners (7-day window) and product page visitors (14-day window, excluding cart adds). Testing Advantage+ Shopping Campaigns against a manually structured conversion campaign showed ASC outperforming by 18% on ROAS across 6 weeks.
Tracking: implemented Meta Pixel alongside CAPI through a server-side GTM container. Attributed conversions increased by approximately 23% compared to pixel-only, recovering signal lost post-iOS 14. UGC-style video creative outperformed polished brand assets by 31% on ROAS. Creatives were refreshed every 3 weeks once frequency crossed 6 or CTR dropped by 20%.
Q28. How did you implement a paid ads strategy across multi-channel campaigns?
My multi-channel approach follows five principles:
Channel-to-funnel mapping: YouTube and Meta Reels own awareness; Google Display and Meta Feed own consideration; Google Search and Meta Conversion campaigns own the conversion stage.
Unified tracking: GA4 as the primary data source. Consistent UTM naming conventions across every paid URL — documented and enforced across the team. Google Ads and Meta both linked to GA4 for cross-channel attribution.
Audience synchronisation: CRM customer lists uploaded to both Google (Customer Match) and Meta (Custom Audience). Consistent exclusion logic applied across all platforms — all-time converters excluded from every acquisition campaign.
Budget allocation: Initial split of 60% Search / 25% Meta / 15% YouTube, reviewed monthly against GA4’s multi-touch attribution data. Budget shifted toward channels showing the highest cross-channel contribution, not just last-click ROAS.
Creative adaptation: Same core message and offer across all channels. Format adapted per platform — 6-second bumpers for YouTube, 15-second vertical Reels for Meta, responsive text-based RSAs for Search. Creative concepts are tested quarterly; executions are refreshed monthly on high-frequency channels.
Q29. What are Meta Pixel and Conversion API, and how have you implemented them?
Meta Pixel is a browser-side JavaScript tag that fires events — PageView, ViewContent, AddToCart, InitiateCheckout, Purchase, Lead — when users interact with your website, sending this data to Meta’s servers. It underpins website custom audiences, conversion tracking, and ad optimisation.
The iOS 14+ problem: Apple’s App Tracking Transparency framework caused most iOS users to opt out of cross-app tracking. The result is 20–40% data loss at the pixel level — conversions go unmeasured, audiences shrink, and optimisation signals degrade.
Conversion API (CAPI) resolves this by sending events server-side, directly from your web server or CRM to Meta — bypassing browser restrictions, ad blockers, and iOS privacy settings entirely.
The correct setup is Pixel + CAPI running simultaneously, with event deduplication via a shared event_id parameter. When both sources report the same event with matching IDs, Meta counts it once — no double-counting.
My implementation:
- Deployed CAPI via a server-side GTM container hosted on Google Cloud Run
- Used Order ID as the
event_idfor purchases; form submission ID for leads - Passed hashed customer data parameters (email, phone, first/last name, location) to improve match accuracy
- Monitored Event Match Quality (EMQ) score in Meta Events Manager — achieved 7.2/10 (industry target is 7+)
- Configured Aggregated Event Measurement priority order for iOS traffic: Purchase → Lead → AddToCart → ViewContent
- Outcome: 25–30% recovery in attributed conversions and measurably improved lookalike audience quality
What most candidates miss: The Event Match Quality score in Events Manager. A higher EMQ score means better profile matching, which directly improves conversion reporting accuracy and algorithmic optimisation. Citing this metric in an interview signals genuine implementation experience.