Every media planner knows the workflow: pull data from one place, paste it into a spreadsheet, rebuild the model you built last quarter, reconcile the numbers that don’t match, and hope the plan holds once the budget starts spending. It’s a slow process, run by hand, on numbers that often don’t hold up.
That’s the problem that Amazon Ads’ Willian Correa teed up in a joint session with Pacvue’s Jocelyn Jeffries at this year’s unBoxed. In the session, Media Planning: From Audience Signals to Planning to Performance with Amazon Ads API, Willian and Jocelyn covered the challenges brands face when it comes to answering questions around audience signals, reach forecasting, and deduplication – and what changes when those questions become easier to answer by moving from something estimated by hand to something available by API.
Here’s what was covered and what it means for brands and agencies.
Amazon’s Media Planning API Suite
One of the hardest numbers to get right in media planning is reach. If the same shopper shows up in a display campaign, an audio campaign, and a streaming TV campaign, a plan that simply adds up reach across all three will overstate how many people it’s actually reaching. The practical risk of not knowing whether a plan’s reach number is real is straightforward and, oftentimes, expensive. Budget gets allocated based on a number that’s partly fictional, and the plan doesn’t get corrected until performance data comes in after the campaign is already live. By then, the money’s spent either way.
Getting a reliable answer on reach (a deduplicated number, not the summed one) has historically required a lot of manual reconciliation. But Amazon Ads now offers reach forecasting, performance forecasting, and cross-channel deduplication through its APIs, meaning advertisers can get a forecast of who a plan will reach, and how much of that reach is unique, before committing budget to it. For anyone planning across sponsored ads, display, audio, video, and streaming TV at once, this addresses something that’s been genuinely hard to do by hand.
How Pacvue is Putting Amazon’s APIs to Work
Even reconciling reach by hand for a single brand, for a single quarter, is real work. For agencies, doing it for a full client roster, every planning cycle, creates an even bigger bottleneck.
Pacvue’s planning workflow is built to remove that bottleneck by taking Amazon’s forecasting and deduplication APIs and turning them into something a media team can run inside a real planning cycle. This cuts down the manual work of pulling disparate data points together and calculating expected outcomes, making it so that the breadth of data Amazon and Pacvue have access to translates into faster answers, not more spreadsheets.
Real-World Applications
During the session, Jocelyn walked through two real client stories that show what this looks like in practice – including one where the answer turned out to be spending less.
The first was a beauty brand with a best-selling product that was, by every conventional measure, still working. But, as Jocelyn pointed out, reliable performance and incremental performance aren’t the same question:
- The Situation: The real question was whether the product’s advertising was still bringing in customers who wouldn’t have otherwise found the brand, or whether it was background noise to people who were buying regardless. To make the most effective use of their budget, the brand needed to know:
- How much additional demand can the best seller capture?
- How much of the available conversion opportunity has already been reached?
- When does incremental investment begin to diminish?
- Should our next dollar go to the best seller or another product?
- The Solution: Using reach forecasting and historic reach data, the team mapped the product’s entire addressable market and measured how much of it the brand had already reached.
- The Results: The results showed that they’d passed the point of diminishing returns. Rather than keep funding a channel that had already captured its available audience, the brand reallocated 15% of that product’s budget to support a new product launch planned for 2027.
The second story Jocelyn shared showed how impactful reliable forecasting is when planning for periods with no room for error:
- The Situation: A household consumables brand generates 40% of its annual revenue in the fourth quarter alone, which means there’s effectively no recovering from a Q4 that goes wrong. To ensure success, the brand needed to answer:
- Where can our budget have the greatest impact?
- How do we build an effective full-funnel plan across display, video, retargeting, and search?
- How do we confidently understand expected reach and performance before committing budget?
- The Solution: The team used Amazon reach, performance, and deduplication forecasting together to build a full media plan across 11 different ad types ahead of the quarter, informed by the brand’s historical CPCs, impressions, and Q4 seasonality.
- The Results: This model gave the team a significant head start on a plan that would have taken considerably longer to pull together by hand, resulting in 90% time savings in building out comprehensive Q4 recommendations. The plan included estimated and incremental reach and performance for each tactic and identified multiple budget opportunities before campaigns launched.
Both scenarios exemplify the same shift that Willian pointed to earlier in the session: moving optimization ahead of launch doesn’t just make it faster; it changes what kind of decision gets made in the first place. It’s the difference between a correction that gets made after the money’s been spent, and a proactive choice made before a campaign launches.
The Takeaway for Brands and Agencies
Any campaign plan is only as good as the numbers behind it, and getting those numbers right has usually meant doing the work by hand, after the fact. With Amazon’s forecasting and deduplication APIs built into Pacvue’s planning workflow, that check can now happen first, enabling teams to do everything from catching a reach number that was never as big as it looked to recognizing when a top performing product has quietly stopped driving incremental growth. Plans that used to get corrected in hindsight can now be built on a reliable answer from the start.
For a brand running media in-house, that means fewer surprises showing up in a quarterly review. For an agency, it means planning work that used to eat days per client, per quarter, can happen in a fraction of the time — freeing up hours for the work that requires a team’s strategic judgment.
To see this in action inside Pacvue, book a demo.
Dig into everything that was announced at Amazon unBoxed at our live webinar and AMA, unBoxed unPacked, on October 7. Register here.
