When Gary Alsakar left Wall Street to start MARLEY NYC, he brought something unusual to influencer marketing: the analytical rigor of a hedge fund.
His agency connects brands like PetSmart and Blue Buffalo with a curated network of over 1,200 influencers across niches. It started with 300 elite dog accounts for pet brands, then expanded into cats, foodie moms, and makeup. But rather than relying on gut instinct or follower counts alone, Gary applies statistical modeling to predict which creators will actually deliver results.
The foundation? Data scraped twice monthly from Instagram and TikTok using Apify.
How the pipeline works
MARLEY NYC uses Apify’s Instagram scrapers and TikTok scrapers to pull engagement data for every influencer in their network. Each scrape captures followers, recent posts, likes, comments, and other engagement metrics.
The team runs scrapes on the 15th and at month's end, keeping their proprietary algorithms fed with fresh data. While they occasionally run separate scrapers for post data versus profile metrics, the workflow stays stable: "If the accuracy is good, we're happy. Occasionally, we have to rerun a scrape, but 95% of the time, it does exactly what we want."
From there, the scraped data feeds into internal models that score influencer quality based on engagement calculations Gary developed himself. They remove outliers, analyze comment length, and measure audience authenticity.
Beyond vanity metrics
What sets MARLEY NYC apart isn't just the data. It's how Gary interprets it. Most engagement calculators focus on surface-level numbers. Gary applies statistical deviation and text analysis to measure whether engagement is genuine.
The team prioritizes comment quality over follower size. "Ten emojis don't mean good engagement," Gary says. They look at character count, content depth, and post frequency. A micro-influencer with 50,000 followers but meaningful interactions can outperform a celebrity with millions.
This approach lets MARLEY NYC forecast an influencer's future performance and offer brands statistical probabilities for campaign outcomes. Social data becomes predictive intelligence.
Why data alone isn't enough
Gary sees the future of influencer marketing as a mix of automation and authenticity. AI can replicate formulas and automate analysis. It can't replace what happens after the data is collected.
Every influencer in MARLEY NYC's network signs a contract and joins a private community where they exchange insights and opportunities. This relational infrastructure ensures trust and exclusivity, something no data model alone can achieve.
"AI can organize data, but it can't call an influencer and say, 'Hey, I need you to pick this brand.'"
Want to build your own influencer tracking system? Start with Instagram Profile Scraper or TikTok Scraper and connect them to your own analytics stack.
