Audience Targeting Best Practices: Interest, Intent and Location

What is audience targeting and why does it matter?

Audience targeting is the practice of choosing who sees your ads, using signals such as interests, recent search behaviour, location, demographics and your own customer data. Done well, it directs budget to people likely to care, which lifts engagement, conversion rates and return on investment.

Poor targeting does the opposite. It wastes spend on uninterested people and can cause audience fatigue when the same irrelevant ad appears repeatedly, which hurts how people see your brand.

Targeting also shapes creative performance. A message that reaches a relevant audience feels personal, so engagement rises across platforms. Treat audience design as a strategic decision, not a setting you tick off at launch.

What are the core targeting types?

Five types cover most campaigns. Each suits a different objective.

Effective campaigns often combine several types.

How do Meta, Google, LinkedIn and DSPs differ for targeting?

Each platform has a distinct strength, so match the platform to the signal you need.

• Meta: strong interest and CRM targeting across Facebook, Instagram and Messenger, good demographic and behavioural segmentation, and lookalike modelling to find people similar to your best customers.
• Google: intent via keywords and in-market segments across Search, Display and YouTube. Its search intent data is something other platforms cannot match.
• LinkedIn: job-based and firmographic targeting, including company size, industry, job function and seniority. It is the strongest option for B2B.
• DSPs: contextual, third-party and geographic layering across the open web, plus frequency capping, dayparting and cross-device targeting.

Contextual targeting, available in DSPs and native platforms, places ads on pages with relevant themes rather than tracking behaviour. It becomes more useful as privacy restrictions grow, and suits sensitive sectors such as finance and health. A mortgage ad beside content on home buying tips is a typical example.

How do you use interest, intent and location targeting?

Interest. Use Meta's predefined interests, such as parenting, sports or travel. Test broad segments first so the algorithm has enough data to find high-value users. Add exclusions, such as past customers, and stack lookalikes (for example 1%, 3% and 5%) to scale while keeping quality.

Intent. Google in-market segments reach people actively researching a purchase. Custom intent lets you build audiences from keywords and URLs taken from GA4 or search terms. DV360 adds third-party intent segments. Align your message to bottom-funnel behaviour with a clear offer and call to action.

Location. Target by country, region, postcode or a radius around a site, and adjust bids by regional performance. Use location extensions to drive foot traffic. Customise creative by city or island, for example Auckland versus Otago, and combine targeting with event calendars such as weather, sport and season. Exclude areas that consistently underperform, and test different radius sizes around high-value locations.

How do you use first-party data and lookalike audiences?

First-party data reduces reliance on third-party cookies. Upload hashed email or phone data to Google, Meta, LinkedIn and other platforms that offer customer matching. The guide notes match rates typically run from 50% to 70%, depending on data quality.

Segment customers by lifecycle stage: new leads, active customers, high-value customers, dormant accounts and churned customers. Refresh lists every 30 to 60 days, because stale data wastes spend. Use the data twice: for retargeting existing customers and as the seed for lookalikes.

For lookalikes, follow four rules:

• Base them on converters, not general leads. A seed of at least 1,000 users is recommended.
• Test sizes. Around 1% gives a close match, while 5% gives scale with less precision.
• Check overlap between segments and refresh seed audiences monthly.
• Build them per platform, as each uses different data signals, and use them for prospecting with broader brand messaging.

How do audience stacking, exclusions and retargeting work?

Stacking layers signals to build a tight segment, for example in-market for SUVs, men aged 35 to 44, in Christchurch, high income and weekend browsers. Too many layers reduce scale and raise costs, so check audience size and overlap, test combinations, and measure whether the stack beats broader targeting.

Exclusions protect efficiency. Suppress converters from prospecting, exclude low-quality or bounced visitors, and keep CRM lists out of cold traffic campaigns. Document every exclusion so overlap errors are easy to trace.

Retargeting works best when you:

• Segment by recency: 7 days for high intent, 14 days for medium intent with possible offers, 30 days for lighter brand touchpoints.
• Sequence messages from awareness to product benefits to a strong call to action.
• Rotate creative every two to three weeks.
• Cap frequency, for example 2 to 3 impressions a day and 12 to 15 a week, adjusted by funnel position.
• Segment behaviour such as page views, time on site and video watch depth (25%, 50%, 75%).

What do audience segments look like in New Zealand?

Segments should reflect how people in each sector actually research and buy. The guide suggests these examples:

• Tertiary: site visitors by programme viewed, course guide downloaders, campus event attendees, international enquiries by country.
• Retail: cart abandoners within 7 days, repeat purchasers by category, loyalty tiers, seasonal shoppers.
• Health: people searching for local clinics or treatment information, prescription refill visitors, health insurance researchers.
• Finance: mortgage calculator downloaders, KiwiSaver comparison users, first-home buyer content readers.
• Tourism: domestic travellers researching regions, international visitors by country, adventure activity researchers.
• Agriculture: farmers researching equipment, seasonal supply purchasers, rural property searchers.

Tailor messaging to each segment's needs and stage in the customer journey.

How do you measure audience performance and avoid common pitfalls?

Measure each segment separately, then compare.

• CTR and CPC by segment show which audiences engage.
• CPA and conversion rate show which deliver return.
• Impression share and overlap reveal where you bid against yourself.
• Holdout tests, where a control group does not see the campaign, show true incremental lift.
• Frequency, recency and cross-channel comparisons guide budget allocation.

Four pitfalls recur:

• Too many small segments, which dilute reach and produce statistically weak data.
• Lists that are never refreshed or rotated.
• Over-targeting, which shrinks inventory and pushes CPMs up.
• No testing strategy across segments, which leaves decisions to assumption.

Frequently asked questions

Written by the KBR team, an independent media and digital agency in Auckland and Christchurch.