Performance Marketing Portfolio
Artem Liholitov
7 years driving growth
Selected case studies: 2023–2025
Dynamic ad customizers to capture high-intent mortgage seekers
Home financing searchers are highly rate-sensitive. They want to know today's numbers, not generic ad copy. Generic RSAs were leaving intent on the table.
Live ad example: date (09.11.2025) and current rate (3,35%) pulled dynamically into the headline via ad customizers
I identified that users searching for mortgage rates were responding to recency signals. They wanted current data, not generic messaging. I built an ad customizer setup connected to live data imports, pushing the current date, today's interest rate, and relevant loan amounts directly into RSA headlines in real time. This made ads immediately feel relevant and trustworthy compared to any static competitor ad. Google heavily favoured these ads in the auction, as the extreme relevance of the dynamic content drove stronger ad rank, leading to a compounding effect across every key metric: more impressions, higher CTR, better CVR, and a lower CPL.
Key insight: Ad relevance in finance is not just about keywords, it is about timeliness. Injecting live data signals (rate, date, loan amount) removed the mental gap between what a user was wondering and what the ad answered, driving quality score, user trust, and Google's own preference for the ad in the auction.
Rebuilding Meta campaigns in a restricted finance niche, cutting CPL by ~50%
Inherited underperforming Meta campaigns from a departing colleague. Finance is a restricted category on Meta, with limited targeting options, strict ad policies, and a skeptical audience that scrolls past anything that feels like a bank ad.
I rebuilt the campaigns from scratch rather than patching what existed. The core lever was creative: I produced short-form video ads with copy engineered to immediately address the specific pain point of the target audience, getting them to stop scrolling within the first two seconds. Alongside this, I implemented proper conversion tracking to feed Meta's algorithm the right signals, allowing smart bidding to find quality leads rather than just volume. Budget remained unchanged throughout.
Key insight: In restricted niches, creative quality is the main variable you can control. The right video and copy combination acts as pre-qualification. It reaches the right audience and eliminates waste by disqualifying the wrong one in the first few seconds.
Scaling beyond Search with Performance Max, incremental volume at half the CPC
Search campaigns were optimized and efficient but structurally capped. The CPA bid strategy had reached its ceiling and could not scale further without deteriorating efficiency. Growth required a new channel, not more budget in the same one.
I identified Performance Max as the vehicle to break the ceiling. Rather than migrating existing campaigns, I built PMax from scratch, investing time in defining strong audience signals and developing ad creatives across formats suited to the channel's multi-surface nature. PMax reached users across Display, YouTube, Discover, and Gmail at a CPC 50% cheaper than Search. While CVR was naturally lower than Search (a known trade-off at the top of the funnel), the volume of traffic more than compensated, delivering incremental conversions at a stable CPA. The result was additive growth on top of the existing Search base, not cannibalisation.
Key insight: PMax is not a Search replacement, it is a volume engine. Lower CVR is expected and acceptable when CPC is 50% cheaper and traffic scales significantly. The key is setting the right audience signals upfront so the algorithm finds users who convert, not just users who click.
These three cases are the ones I can speak to with precise, verified numbers. Over 7 years across agency and in-house roles, there are many more campaigns, tests, and results behind them, across additional channels, markets, and objectives. Happy to discuss any of it in more depth.