The company had been building something genuinely different. A location AI that could analyze user behavior, identify high-value placement opportunities in real time, and guide bidding decisions with a precision most ad networks simply do not have. That was the product. That was the competitive edge.
The problem was not the AI. The problem was everything around it.
Their previous DSP was creating friction at every stage. Retargeting was limited. Campaign pacing was uneven. When the team needed to adjust the platform to match their workflow, the tools were missing. The AI was working. The platform was falling behind.
And that was exactly the moment when Epom came in.
When the Technology Is Ready but the Platform Isn't
The company came to Epom not because its fundamentals were broken. They came because their fundamentals were strong and their platform was holding them back.
Three problems were identified early in the discovery process:
- Missed retargeting opportunities. The platform could not segment users effectively. Audiences were reached once and then lost.
- Uneven campaign delivery. Budgets were spent inconsistently across the campaign period, creating spikes and gaps that eroded client ROI.
- No white-label capability. The platform could not be branded, which limited how the company presented itself to advertisers and made client onboarding harder.
“When we looked at the setup, the AI component was genuinely impressive. But it was operating within a DSP that slowed it down. The retargeting gaps alone were a significant drag on what the platform should have been able to deliver.”
The assessment was clear. The technology deserved better infrastructure.
Why They Needed More Than a Platform Switch
Switching DSPs is not a small decision. For a company whose product is built on location AI and whose clients depend on precise, consistent delivery, the risk of disruption was real.
What the team needed was a provider that could match the specificity of what they had already built. That meant a short list of hard requirements:
Why They Choose Epom:
- Integrate with existing AI bidding logic without rebuilding from scratch
- Offer segment-based retargeting out of the box
- Provide granular campaign controls: even pacing and timezone targeting
- Support white-label capabilities for client-facing deployments
- Move from first call to live campaigns in under six weeks
“We do not ask clients to adapt to the platform. We adapt the platform to the client. That principle is what drove how we approached this implementation.”
The company did not look at other providers after the first detailed call. The entry point was right. The technical flexibility was there. The timeline was realistic.
Four Changes, One Direction
The implementation was structured around four specific interventions, each addressing an identified gap. The timeline from first call to live campaigns was under six weeks.
| Timeline | Action |
|---|---|
| Week 1-2 | Discovery and platform audit. Mapping of existing AI logic to Epom's bidding API. |
| Week 3 | AI integration completed. Location AI connected to Epom DSP as part of the bidding process. |
| Week 4 | Even Pacing activated. Campaign delivery rules configured for all active client campaigns. |
| Week 5 | Retargeting by segments deployed. Audience collection logic configured and tested. |
| Week 6 | Timezone targeting activated. First full campaign cycle completed with all four features live. |
Each change addressed one gap. Together, they removed every constraint that had been holding back the AI layer.
“The first thing we needed to fix was delivery consistency. Once pacing became predictable, the team had cleaner reports, steadier client conversations, and less wasted budget.”
How the Integration Actually Worked
The most technically significant change was connecting the company's existing location AI to Epom DSP's bidding layer. This was not a replacement of Epom's optimization logic. The AI worked alongside it.
In practice, the AI analyzed historical placement performance across the company's client campaigns and produced real-time signals about which placements to prioritize and at what bid level. Epom DSP received those signals and acted on them within the auction window.
| Layer | What It Does |
|---|---|
| Client AI Layer | Analyzes historical placement performance. Outputs bid recommendations in real time. |
| Epom DSP Bidding API | Receives AI signals and executes bids. Connects to 50+ SSPs for inventory access. |
| Epom Pacing Controller | Distributes impressions, clicks, and budget evenly across the campaign period. |
| Segment Retargeting Tool | Collects users automatically by behavior. Re-engages audiences that showed intent. |
| Timezone Targeting Rules | Filters delivery by local time. Campaigns reach users at the moment clients specify. |
| Impression Delivery | Ads served to the right audience, at the right time, on the right placements. |
The retargeting layer added a second dimension. Users who interacted with a campaign were automatically collected into segments. The platform began re-engaging those audiences without manual configuration from the team. This replaced the fragmented approach the company had previously used across multiple tools.
Timezone targeting was the final piece. Clients in time-sensitive verticals needed their impressions to land at specific hours in specific markets. Epom's timezone targeting setup reduced that to three steps.
Results in Numbers
The results arrived within the first full campaign cycle after all four changes were live. Every metric the company had been targeting moved in the right direction.
| Metric | Before Epom | With Epom | Change |
|---|---|---|---|
| Average CTR | 0.8% | 1.1% | +37% |
| Conversion Rate | Baseline | +25% vs baseline | +25% |
| New High-Value Clients | Irregular | 3 per quarter | Consistent |
| Ad Spend Efficiency | Uneven delivery | 20% waste reduced | -20% |
The CTR movement from 0.8% to 1.1% was the figure the company could point to in client conversations. The 20% reduction in ad spend waste was the figure that changed the economics of running campaigns at scale.
Timezone targeting was the final piece. Clients in time-sensitive verticals needed their impressions to land at specific hours in specific markets. Epom's timezone targeting setup reduced that to three steps.
“The numbers moved in the right direction across every metric we tracked. That does not happen often when you switch platforms. Usually, something breaks. Nothing broke.”
Three new high-value clients joined in the first quarter following the implementation. A better platform performance record is a sales asset. The team used it as one.
What This Case Points To
The US AI Geo Location Ad Provider is continuing to grow on Epom DSP. The team is evaluating expanded white-label capabilities as their client base scales. The direction is clear.
The broader point is straightforward. A strong proprietary technology layer requires infrastructure that matches its ambition. The AI was working. Epom gave it a platform that could keep up.
