How I Built an AI-Powered Influencer Campaign Tracker and Uncovered £6,000+ in Hidden Revenue

# How I Built an AI-Powered Influencer Campaign Tracker and Uncovered £6,000+ in Hidden Revenue ## Turning a Marketing Problem into a Scalable Internal Tool At my workplace, I manage gifted-product influencer campaigns across YouTube, Instagram and TikTok. On the surface, the process seems straightforward: send products to creators, they make content, and you track the results. The reality was far more complicated. We had 19 active creators, more than 80 pieces of content, dozens of product SKUs, discount codes generating revenue, and campaign data spread across Google Sheets, browser tabs and email threads. There was no centralised view of campaign performance, no reliable way to track which products had actually been featured, and no clear connection between content and revenue. Most influencer marketing platforms are designed around paid sponsorships and fixed campaign fees. They don't handle the day-to-day realities of gifted campaigns particularly well. I needed to know: * Which products each creator had covered * Which products were still outstanding * How content was performing over time * Which discount codes and links were generating revenue * Which creators were delivering the strongest results The existing workflow involved manually checking social profiles, updating spreadsheets and pulling data from multiple sources whenever a report was required. It wasn't scalable and was consuming hours every week. So I decided to build my own solution. ## Building an Influencer Campaign Tracker with AI Using AI as a development partner, I built a browser-based Influencer Campaign Tracker from scratch. The application is entirely self-contained, requiring no traditional server infrastructure or database management. It runs directly in the browser and centralises every aspect of campaign management into a single platform. Each influencer profile contains: * Creator details and social handles * Product SKUs and gifted values * Content links and platform information * Historical performance snapshots * UTM and affiliate tracking * Discount code performance * Revenue attribution * Communication and outreach tools The goal was simple: create a single source of truth for every influencer campaign. ## Historical Performance Tracking One of the most useful features became historical snapshot tracking. Most social media tools show performance at a single moment in time. That's useful, but it doesn't tell you how a piece of content is performing over weeks or months. I wanted to see growth patterns. Instead of simply knowing that a video has 400,000 views today, I can now see that it achieved: * 12,000 views in its first week * 85,000 views after three weeks * Continued growth over several months This allows me to identify content with long-term performance potential rather than relying solely on vanity metrics. For influencer campaigns, understanding growth curves is often more valuable than looking at a single view count. ## Automating Analytics with APIs and Apify The biggest time saver came from automation. I integrated the YouTube Data API alongside multiple Apify scripts to automatically pull performance metrics from YouTube, Instagram and TikTok. With one click, the tracker updates: * Views * Likes * Comments * Engagement metrics * Historical snapshots Before building the tool, gathering this information meant manually checking individual posts and copying figures into spreadsheets. Now, the entire process is automated. Connecting Apify scripts directly into the tracker has saved hours of repetitive reporting work and eliminated one of the most time-consuming parts of managing influencer campaigns. Instead of spending time collecting data, I can spend time analysing it and making decisions. ## Solving Technical Scaling Challenges Building the application was one challenge. Making it scale was another entirely. The initial version stored campaign data in localStorage and synchronised information with Google Sheets using Apps Script. It worked perfectly at small scale. Then the data started growing. As more influencers, posts and historical snapshots were added, I began encountering: * Browser URL length limitations * Google Apps Script quotas * Google Sheets cell limits * localStorage storage restrictions At one point, new information simply stopped saving. Statistics disappeared after refreshes and settings repeatedly needed to be re-entered. Rather than compromising functionality, I redesigned the architecture. The application was migrated to IndexedDB, allowing significantly larger datasets to be stored locally, while the Google Sheets synchronisation system was rebuilt to safely handle large amounts of data. Following the redesign, the application successfully handled thousands of historical snapshots without performance issues. The experience forced me to think beyond simply writing code and instead focus on designing systems that remain reliable as data and complexity increase. ## Discovering More Than £6,000 in Previously Untracked Revenue The most commercially significant feature became attribution tracking. Each influencer profile contains dedicated sections for: * UTM link performance * Affiliate revenue * Coupon code performance * Orders and conversions * Revenue attribution Before implementing this system, campaign success was primarily measured through engagement metrics. The connection between content and revenue was largely invisible. As attribution data was added, the picture changed dramatically. The tracker uncovered more than £6,000 in directly attributable revenue that had previously gone unmeasured. Individual creators generated thousands of pounds in revenue through discount codes and affiliate activity that otherwise would have been difficult to identify and report. The project fundamentally changed how campaign success was measured. Instead of saying, "This campaign seems to be doing well," I could now demonstrate exactly how much revenue individual creators were generating. ## Improving Campaign Operations The project also solved several operational challenges. One recurring issue was identifying which influencers still owed content for products they had received. Managing this manually across multiple creators and product lines was becoming increasingly difficult. I built a communications system that provides an instant overview of: * Fully covered products * Partially covered products * Outstanding products * Relevant content links * Automated follow-up templates The application can generate draft outreach emails using dynamic variables, significantly reducing administration time. Additional features include: * Asset repurposing management * Content permissions tracking * Automated reporting * Dashboard summaries * Exportable analytics data * Public campaign views for stakeholders The result is a workflow that is considerably faster, more organised and far easier to scale. ## The Results Since implementation, the tracker has managed: * Over 2.6 million impressions * More than 80 pieces of content * Over £12,000 in gifted product value * More than £6,000 in attributable revenue * A campaign engagement rate exceeding 4% * Comprehensive historical performance data across multiple creators and platforms Most importantly, it transformed campaign management from a process driven largely by intuition into one supported by measurable data and automated reporting. ## What This Project Taught Me About AI and Problem Solving This project completely changed my perspective on AI-assisted development. AI didn't simply help write code. It became a collaborative problem-solving tool that helped diagnose production issues, test solutions, challenge assumptions and accelerate development. More importantly, the project demonstrated the value of combining marketing knowledge with technical problem solving. I identified a genuine operational problem, designed a solution around real business requirements, automated time-consuming processes and built a system that improved reporting, uncovered revenue and saved hours of administrative work every week. What began as an attempt to organise influencer campaign data evolved into a production tool that now manages every stage of the workflow, from product dispatch and content tracking through to attribution and reporting. For me, the biggest lesson was simple: some of the most valuable projects come from solving your own problems first. When you build tools around real operational pain points, the result is often something far more impactful than you originally intended.