How to read ad strategy through administrative-dong living population

Seeing the people outside your ad account

Inside an ad account, you live with clicks, conversions, and ROAS. But for brick-and-mortar businesses or OOH, the first questions come from outside the account: When is this area actually busy? Does a weekday lunch behave differently from a weekend evening? Can you really run the same creative in an area packed with 20-somethings as in one full of people in their 40s?

Resident-registration counts alone can’t answer these questions. Once you factor in office workers, students commuting in and out, shopping and dining visitors, and nighttime foot traffic, the population on file by home address looks nothing like what’s actually on the ground. That’s why Neuro added Seoul’s administrative-dong–level living population data — not to surface one more public dataset, but to read ad performance against the local rhythm of the day.

What the data looks like inside Neuro

The dataset we connected here is the administrative-dong–level Seoul Living Population (Korean nationals) from the Seoul Open Data Plaza. Per the official description, it estimates the population present in a given area at a given time by combining Seoul city public data with telecom data. In Neuro, we structured it so you can query by date, daypart, administrative-dong code, city/gu and dong name, total living population, and living population broken out by gender and age group.

When you name a region in natural language, Neuro first searches candidate administrative dongs to lock in the exact one, then queries the living population table. A request like “Show me the weekend-evening living population for 20-somethings in Seongsu-dong,” for instance, has to be split into place-name resolution and time conditions handled separately. That’s how we cut the risk of reading the wrong area because of similar dong names or vague district-level phrasing.

Putting it to work in franchise F&B

In franchise F&B, regional factors sit heavily between ad performance and actual store visits. Even with the same coffee-coupon campaign, a weekday morning in an office-heavy dong, an afternoon near a university, and a weekend evening in a residential area should be treated as separate campaigns. With living population layered in, you can tune messaging around “the age groups that actually swell in the dongs around this store at specific dayparts” rather than “every 20-something in Seoul.”

• In areas with a sharp lunch peak, lead with messaging around quick table turnover, set menus, and mobile ordering.
• In areas where living population climbs after work, look at delivery, takeout, and dinner promotions together.
• In trade areas with big weekend swings, don’t cut budget off weekday-average ROAS alone — re-split it by day of week.
• Before opening a new franchise location, compare daypart living population across candidate dongs against the sales patterns of existing stores.

The point is not to use living population as a targeting option. Ad-platform targeting and living population are different datasets. In Neuro, living population is an input for “what regional and daypart hypotheses should we form,” while the actual performance read has to come from ad metrics, store sales, coupon redemptions, and order data together.

Putting it to work in OOH and local media

OOH is bound to place and time even more tightly than online advertising. When you evaluate subway, bus, billboard, or elevator media, “lots of foot traffic” isn’t enough to go on. You need to see which dongs run strong by day versus by night, how wide the weekday-versus-weekend gap is, and which dayparts make your brand’s target age groups stand out the most.

With this data in Neuro, you don’t jump straight to a buy plan on broad units like “all of Gangnam-gu” when shortlisting OOH placements — you can narrow candidates down to the administrative dong and daypart. The bigger the media budget, the more you should commit a hypothesis to writing before you buy. Logging why you picked this area, which population signals you saw at which dayparts, and how you expect it to split roles with your online campaigns is what lets you sharpen the next placement.

How to ask Neuro

A good prompt spells out region, period, daypart, and comparison criteria all at once. Instead of “How’s Seongsu-dong doing?”, a far better query is: “Across the Seongsu-dong cluster of administrative dongs in Seongdong-gu, show the weekend-afternoon living population trend for ages 20–34 over the last four weeks, and recommend dayparts for an F&B pop-up campaign.” The agent confirms the candidate regions first, queries the living population, and — if ad performance or store order data is connected — compares it side by side over the same period.

1. Confirm the administrative-dong name first. If several dongs share a name or you’ve only given the district, Neuro surfaces candidates and asks you to confirm.
2. Set the period and dayparts. Decide whether to look at daily averages, split weekday from weekend, or align with a specific campaign window.
3. Read total living population alongside the gender and age breakdown. Even when the overall volume is high, you need a different play if your target segment is thin.
4. Tie it to other data — ad performance, store sales, coupon redemptions, search trends. Don’t lock in budget decisions on living population alone.
5. Finally, write down your execution hypothesis. Logging which messages you’ll test in which regions and dayparts is what makes the next analysis easier.

What not to read into it

Living population is an estimate. The official documentation itself notes that it can diverge from reality for certain areas or moments in time. The public data doesn’t track individuals, and it’s not a real-time sensor telling you how many people are standing in front of your store right now. So don’t read it as purchase intent, probability of a visit, or guaranteed ad impressions.

Drawing conclusions about an area from a single day of data is especially risky. Weather, events, public holidays, school calendars, large rallies, and concerts all creep in easily. When you use living population in Neuro, at a minimum set comparison periods, compare like-for-like days of the week, and read it alongside campaign performance. More data doesn’t automatically make a decision more accurate.

Prompts you can use right away

• Compare weekday lunch and dinner living population across three candidate administrative dongs for a new store location.
• Find the Seoul administrative dongs with strong living population for ages 20–34 on weekend afternoons this month, and write them up as F&B pop-up candidates.
• Compare day-versus-night living population across our OOH candidate areas, and suggest how each could support our online campaigns.
• Show the living population trend for the administrative dongs around a specific store side by side with how ad ROAS moved over the same period.

What Neuro is after isn’t just explaining the numbers inside your ad accounts well. It’s helping ad operators read the full context — the regions, times, and customer segments where budget actually goes — and ask sharper questions before they execute.