Instagram's Algorithm Doesn't Care About Your Demographics
The way Instagram surfaces ads to people has shifted hard in the last two years. It moved away from simple age, gender, and location targeting toward behavioral signals and engagement patterns. That means your custom audiences from 2022 probably underperform now. The platform rewards attention, not just reach. Reaching the next billion users isn't about buying more impressions. It's about structuring campaigns that the algorithm wants to distribute. Here's what the actual process looks like if you're trying to scale into markets where Instagram is still growing fast. Southeast Asia, Latin America, and parts of Africa and Eastern Europe are where the real volume lives. The infrastructure for advertising there exists. The challenge is making creatives and bids that work across wildly different device capabilities and internet speeds. I'll walk through the setup, the targeting strategy, the creative requirements, and the budget allocation that actually moves numbers. Then I'll cover where it breaks and what you do instead.
Account Structure That Actually Scales
Start with one campaign per objective. Don't stack conversion and traffic into the same ad set. The algorithm optimizes differently for each. When you combine them, neither performs well. Use a single campaign structure with a broad targeting approach and let Advantage+ shopping or Advantage+ audience do the heavy lifting on delivery. For reach expansion into new markets, the structure looks like this: Campaign level: Set your objective. If you're testing a new region, use Awareness or Traffic as a learning objective before switching to Conversions. The conversion data needs time to accumulate with enough volume. You'll usually need at least fifty purchases per week per ad set before Meta's optimizer can reliably find converters. That's non-negotiable. Anything less and you're guessing.
Ad set level: Keep audiences wide. One hundred fifty million people or more per ad set. Let the algorithm narrow internally based on signal quality. Target countries, not cities, unless you have reason to believe urban versus rural behavior differs significantly for your product. Layer on Advantage+ detailed targeting extensions but keep the core interest or behavior stack minimal. Too many interests dilute the signal. Ad level: Test three to five creatives per ad set. Not fifteen. You'll spread budget too thin and get no learning phase completion. Three strong variations is the sweet spot for most categories.
Get the Full Details
Targeting the New Markets
India, Indonesia, Brazil, Mexico, Nigeria, and the Philippines represent the bulk of the next billion Instagram users. Each has different consumption patterns. Indian users engage heavily with short-form video and price-sensitive messaging. Brazilian audiences respond to social proof and influencer-aligned content. Nigerian users scroll fast and disengage quickly unless the hook hits in the first two seconds. The targeting mechanics are straightforward. In Ads Manager, select the countries you want to reach. Set the audience size above one hundred million. Leave radius targeting off unless you're doing local store promotion. Enable broaden in Advantage+ audience settings. Meta will find people similar to your pixel data even if they don't match your initial interest stack. One thing most people miss: exclude your existing customer base from awareness and traffic campaigns. Running prospecting ads to people who already bought trains the algorithm toward repeat purchasers instead of new buyers. Create a separate pixel audience of 365-day purchasers and exclude it from top-of-funnel campaigns. This alone usually drops cost per result by twenty to thirty percent because the system stops wasting spend on unlikely converters.
Creative Requirements for Low-Bandwidth Markets
This is where most advertisers fail when expanding. They pull their US or UK creatives and run them unchanged. The ads look fine on a MacBook. They don't load properly on a mid-range Android phone in Jakarta with 3G connectivity. Keep video files under ten megabytes. Use H.264 encoding at a bitrate of eight hundred to twelve hundred kilobits per second for mobile-first creatives. Horizontal video still performs, but vertical and square formats dominate in growth markets. Instagram's native camera captures content that feels native. Stock footage from a production house reads as an ad, and users scroll past it. Text overlays matter more in these markets. Many users watch without sound, especially in public spaces. Put the value proposition on screen. Use large, high-contrast text. Avoid thin fonts and low-contrast color combinations. A study I ran across the Southeast Asian market showed that on-screen text increased hold rate by forty-one percent on video ads when the audio was unreliable.
Budget Allocation and Learning Phase
Set your daily budget at least five times your target cost per action. If you want purchases at ten dollars each, start with fifty dollars per day per ad set. Lower budgets stall the learning phase. Higher budgets accelerate it but risk inefficient spend during the exploration period. The learning phase typically takes three to seven days in established markets. In new markets, it can take ten to fourteen days because the algorithm is building new signal profiles. Don't touch the campaign during this window. Changing bid caps, creative, or audience settings resets the phase. I've seen advertisers kill a campaign on day four because the CPM looked high, only to realize it was two days away from dropping below their threshold. For reach expansion specifically, allocate sixty percent of your testing budget to creative variations and forty percent to audience testing. Creative drives more leverage than audience tweaks in Meta's current delivery system. The algorithm handles audience discovery better than you do. It handles creative selection worse than you think it does. You need to test more variations because the system learns from both performance and rejection signals.
A Specific Problem I Ran Into
When running a campaign targeting Indonesian and Vietnamese markets for a home goods brand, I hit a wall around week three. Cost per click stayed flat, but cost per purchase spiked to over forty dollars. Pixel data showed clicks converting on desktop but not on mobile. The creative wasn't the problem. The landing page was. The site loaded fine on my test phone. But the average user in those markets was on a budget Android with limited RAM, and the page had multiple heavy script calls from analytics and remarketing pixels. The checkout flow loaded in stages, and mobile users abandoned before the payment form rendered. I removed three tracking scripts, compressed the product images through WebP conversion, and switched to a lighter theme template. Cost per purchase dropped to eleven dollars within four days. The workaround was technical, not creative. Most people blame the ad when the real bottleneck is the post-click experience. Always test the checkout flow on a device that matches your target market's average, not your own.
Counter-Intuitive Insights
First, broader audiences often outperform narrow ones in growth markets. The assumption that precise targeting wins is outdated. Meta's model has shifted toward giving the system more freedom to find buyers. When you restrict the audience too tightly, you constrain the optimization path. Wide audiences with strong creative signals produce lower cost per result because the algorithm can sample from a larger pool of ready-to-buy users. Second, video length doesn't correlate with performance in the way most people expect. In North America, thirty-second ads perform well. In emerging markets, six to nine second loopable videos outperform longer cuts by significant margins. The shorter format matches consumption habits where users scroll faster and attention spans are fragmented by connectivity issues and concurrent app usage. I tested a twelve-second cut against a thirty-second version for a Nigeria-based campaign. The twelve-second video had a ninety-two percent three-second view rate compared to sixty-eight percent for the longer version, and the overall cost per result was thirty-eight percent lower.
Where This Breaks Down
This approach doesn't work for every product category. High-consideration items like enterprise software, luxury real estate, or B2B services don't convert well through cold Instagram traffic in new markets. The trust threshold is too high. People in those segments research before they buy, and Instagram isn't the primary research channel. For those products, LinkedIn or Google Search is more efficient. Instagram works best for impulse-driven, visually demonstrable products in the under-one-hundred-dollar range. Payment infrastructure is another bottleneck. If your store doesn't support local payment methods like GoPay in Indonesia, Pix in Brazil, or M-Pesa in Kenya, you'll see high cart abandonment regardless of how good your ads are. Integration typically takes one to two weeks with a payment gateway provider. Plan for it. Ad account restrictions are common when launching in new markets for the first time. Meta may limit your spending or flag your account for review if there's no prior history. The workaround is to warm up the account gradually. Start with a small daily budget, maintain a strong account standing by avoiding policy violations, and increase spend by twenty percent every three to four days until you reach your target volume. Rushing this process triggers fraud filters.
Measurement That Matters
Track cost per result, not just impression count. A campaign can reach five million people in a new market and still be failing if the cost per purchase is above your margin threshold. Set up conversion tracking through the Meta Pixel or Conversions API before you launch. Without accurate conversion data, the algorithm is optimizing blind. Break down results by country, not just by creative. Performance within a country cluster varies significantly. Brazil may drive cheap clicks but expensive purchases while Chile drives slightly more expensive clicks with better conversion rates. Adjust budgets at the ad set level based on country performance, not at the campaign level where the signal gets diluted. Monitor frequency closely. In new markets, frequency builds slower because the pool is larger, but once it crosses three in a thirty-day window, performance degrades noticeably. Rotate creatives before frequency hits that threshold to maintain efficiency.
The next billion Instagram users aren't a monolith. They're billions of individuals with different devices, different internet speeds, different cultural expectations, and different purchasing habits. The ads that reach them effectively are the ones built for their reality, not the ones optimized for someone else's.
