Halloween 2025: $75K Campaign Yields 6.2x ROAS

Halloween 2025 was tough. We were staring down a fragmented audience and ad costs that just kept climbing. So when we took on a regional costume retailer’s campaign, “Spooky Savings & Thrills,” the goal was simple: own the pre-Halloween market in the Atlanta metro. We set out to prove that smart, data-driven advertising could still pull in a serious return on ad spend, even when everyone’s fighting for attention during the holidays.

The “Spooky Savings & Thrills” campaign ran from September 20th to Halloween night, October 31, 2025. We put a $75,000 budget to work across Meta (Facebook and Instagram), Google Search, and display networks to drive both online sales and foot traffic into the client’s three stores in Alpharetta, Sandy Springs, and Midtown Atlanta. We knew we were up against big online players and all the pop-up shops, so our entire strategy was built on being more local and using their past sales data more intelligently than anyone else.

Strategy: The Hyper-Local & Behavioral Blend

Our strategy was a mix of geographic targeting and behavioral segmentation. Geographically, we drew 10-mile radius circles around the client’s brick-and-mortar stores, hitting zip codes in places like Buckhead, Dunwoody, and Roswell, where we knew there were families with more disposable income. Just as important, we actively excluded areas that had historically burned money in past campaigns with low conversion rates, which instantly made the budget work harder.

Behavioral segmentation was where we really expected to get an edge. We started by building custom audiences from people who’d visited the website and spent over $50 during the previous Halloween season. That data became the foundation for our lookalike audiences on Meta, targeting new users who acted and looked just like their best past customers. We also layered in broader interests like “Halloween costumes” or “cosplay” to fill the top of the funnel, but the real money-maker was our retargeting plan: a full 60% of the budget was dedicated to chasing down users who had viewed products or abandoned their carts.

Looking at the competition, we saw a lot of big, generic national campaigns. We figured our local angle would cut through the noise. It let us tailor ad copy with hyper-local references, like “Perfect for the Candler Park Fall Fest!”, a small detail that builds a ton of relevance and trust with an Atlanta-based shopper. It’s the kind of thing that big national brands just can’t do well.

Creative Approach: Beyond the Spooky Stock Photo

We completely ditched the polished, professional product shots. The creative strategy was all about user-generated content (UGC) and short, punchy video ads. In early September, we ran a small social contest asking people to post photos of last year’s costumes, which gave us a library of authentic content that performed better than anything we could have shot in a studio. We also worked with a handful of Atlanta-based micro-influencers (in the 5k-20k follower range) for costume try-ons and unboxing videos that felt real, not like a slick ad.

The specific ad formats we ran were:

• Instagram Reels: Fast 15-30 second videos showing off costumes with quick cuts and trending audio.
• Facebook Carousel Ads: These let us group costume categories like “Kids’ Costumes” or “Pet Outfits” and link each card directly to the right landing page.
• Google Search Ads: Standard text ads for anyone searching high-intent keywords like “Halloween stores Atlanta” or “costumes near me.”
• Google Display Ads: We placed static and animated banners on sites our target audience already visited, like local Atlanta parenting blogs and event calendars.

To keep things fresh and enable constant testing, we churned out over 50 unique ad creatives. It was a lot of work upfront, but it meant we could spot winning ads and kill the losers fast.

What Worked: Precision Targeting and Authentic Creative

The campaign delivered a strong 6.2x return on ad spend (ROAS), blowing past our 4.0x goal. For every $1 we spent, we brought in $6.20 in revenue. Here’s what drove those numbers:

1. Lookalike Audiences: They were the clear winners. They brought in a cost per conversion (CPC) of $8.50, while our broader interest-based audiences were way higher at $12.30. The conversion rate from lookalikes was 5.8%, a full 22% better than general targeting, which just proves that using your own first-party data is the best way to find new customers.
2. Instagram Reels: Short-form video on Reels averaged a click-through rate (CTR) of 2.8%, easily beating our static images (2.0% CTR) and even our Facebook video ads (2.2% CTR). The fast, authentic feel just worked for parents and younger shoppers looking for quick ideas.
3. Urgency-Driven Ad Copy: Our A/B tests proved that copy with urgent language like “Last Chance for Spooky Savings!” had a 15% lower cost per conversion than generic sale messages. Creating that fear of missing out got people to buy faster.
4. Retargeting Success: The 60% of our budget dedicated to retargeting paid off with an 8.5x ROAS. It’s no surprise that users who had already shown interest were far more likely to buy when we reminded them with ads, especially when we showed them the exact products they’d already looked at.

Overall, we served 12.5 million impressions and maintained an average CTR of 2.1%. Our cost per lead (CPL) landed at $6.10, and with 8,823 total conversions (a mix of online sales and in-store coupon redemptions), our final average cost per conversion was $8.50.

One ad in particular, which featured a real local family in superhero costumes filmed in Piedmont Park, was a standout. That single ad got a 3.5% CTR and drove 12% of all conversions from Instagram. It’s the perfect example of how relatable, local content can destroy a polished, generic ad any day of the week. This lines up with what industry reports are saying, like a recent eMarketer report on 2025 holiday shopping, which noted that consumers want more authentic interactions from brands.

What Didn’t Work & Optimization Steps

Not everything worked right out of the gate. Our first attempt at Google Display ads using broad interest targeting like “Halloween enthusiasts” was a money pit. The CPC was a painful $18.20, and the traffic it sent was low-quality, with conversion rates in the basement.

Optimization steps included:

• Refining Display Network Targeting: We pivoted fast. We killed the broad interest targeting and shifted to placing ads on specific niche websites and apps we knew our audience used. We also built custom intent audiences to target people who had recently searched for our competitors, which dropped our Display Network CPC down to $11.50 within two weeks.
• Ad Creative Refresh: About three weeks in, we saw CTRs start to dip as ad fatigue set in. We swapped in a new batch of UGC and influencer content that was focused on humor and last-minute ideas, which helped bump engagement back up.
• Bid Adjustments: This was a daily grind. We constantly monitored performance, pushing more budget to the ad sets that were working and pulling back or pausing the ones that weren’t. It’s manual work, but it’s where performance marketers earn their keep.
• Landing Page Optimization: After noticing high bounce rates on a few category pages, we worked with the client to make some quick fixes. We added clearer call-to-action buttons and made sure the pages loaded faster on mobile, and those small tweaks made a real difference in conversions.

In-store conversion tracking was another early headache. The unique discount codes we offered weren’t always used. To get a clearer picture, we enabled Google Ads’ store visit conversions. While it’s an estimate, having that data gave us a much better sense of the foot traffic our campaigns were actually driving, which is essential for any retailer with a physical footprint.

The Learning Curve: Future Halloween Campaigns

The “Spooky Savings & Thrills” campaign gave us a ton of good data for the next holiday season. We confirmed that while you need broad reach to an extent, the best ROAS comes from precision targeting, especially lookalikes and aggressive retargeting. Showing how real people use and enjoy the products, through authentic UGC, is what separates you from the competition.

For 2026, we’re already planning to test more interactive ad formats like polls and quizzes in social media stories to get engagement up even higher. We’re also looking at partnering with local community groups on sponsored events, tying our digital ads to things happening in the real world. According to IAB’s 2025 State of Internet Advertising report, that kind of hybrid online-to-offline strategy is what’s driving growth for local businesses right now.

With ad costs continuing to rise, the results here show that a smart strategy with constant, daily analysis and agile optimization still wins. There’s no single trick. The key is consistent, granular effort.

The big takeaway is that you don’t need an astronomical budget to make a huge impact during a peak season. If you get granular with your audience targeting and aren’t afraid to iterate on your creative until you find what works, you can get serious returns.