Daniks.AI vs Manual Amazon PPC: NordFrost Case Study

In the week of May 18–24, 2026, an insulated drinkware brand on Amazon US ran its Sponsored Products and Sponsored Display campaigns almost entirely through Daniks.AI. The AI-managed side delivered 7.77x ROAS at 12.9% ACoS — 149 orders and $3,462 in ad-attributed sales from just $446 in spend. A vestigial manual track ran alongside it with $10 in spend, 15 clicks, and a single order at 2.07x ROAS and 48.3% ACoS. The blended TACoS across the entire account was 8.1% — meaning PPC spend represented just eight cents on every dollar of total revenue.

This is not a balanced side-by-side comparison. This is what Amazon PPC looks like when a brand has handed the bid math almost entirely to an AI agent and the numbers have stabilized.

Ownership disclosure. I am Ekaterina Rubtcova, the founder of Daniks.AI, the AI-native Amazon PPC automation that managed the robot side of this report. I built it for my own listings first — the Daniks cookware brand that reached Top-1 in Germany and is currently Top-20 in the USA. NordFrost is a customer brand running on Daniks.AI, not my brand. They sell insulated stainless steel drinkware on Amazon US — tumblers, water bottles, and travel mugs. The weekly report data below is from their seller account with permission to publish.

What this weekly report shows

The Daniks.AI Amazon App sends automated weekly reports to every connected seller account. This one covers a single week — May 18 through May 24, 2026 — on Amazon US. The report breaks down Sponsored Products and Sponsored Display performance into two tracks:

Robot-managed campaigns: Daniks.AI running campaign creation, keyword harvesting, bid adjustments, negative keyword management, and target-ACoS optimization on autopilot. The brand allocated 98% of its weekly ad budget here.
Human-managed campaigns: a small manual track kept alive as a baseline reference — $10 in spend, 15 clicks, minimal activity. This is not a competitive manual program; it is a canary.

Unlike the previous case studies on this site — Tropeza, Fornel, Basecamp Roasters, StellaHive — this is not a balanced A/B. The brand has already decided: the AI runs the PPC. What this report shows is the steady-state result of that decision.

Key Takeaways

Daniks.AI delivered 7.77x ROAS at 12.9% ACoS — the highest ROAS and lowest ACoS of any published Daniks.AI case study.

CPC was $0.39 on the robot side vs $0.64 on the vestigial manual side — 39% cheaper clicks.

CVR was 12.9% — nearly one in eight clicks converted to an order. The manual side’s 6.7% CVR (1 order from 15 clicks) is directional at best.

TACoS was 8.1% — the brand spent $455 on ads against $5,639 in total revenue. PPC is funding growth without dominating the P&L.

149 orders in a single week from $446 in AI-managed spend. Average cost per order: $2.99.

NordFrost on Amazon US

NordFrost is a US Amazon brand in the insulated drinkware category — stainless steel tumblers, vacuum-insulated water bottles, and travel mugs. Mid-range price point ($18–30 per unit), year-round demand with a summer peak for outdoor drinkware and a Q4 gift-set bump. The product is durable but has natural replacement cycles and strong gifting behavior — someone who likes their tumbler often buys another as a gift.

Insulated drinkware is a competitive Amazon category with well-known brands. Success here requires tight CPC control because margins on a $20–25 tumbler are moderate after referral fees, FBA fulfillment, and the cost of stainless steel manufacturing. The brands that win are not the ones with the most impressions — they are the ones that convert efficiently on the impressions they buy. A 12.9% CVR at $0.39 CPC is the kind of unit economics that lets a mid-size brand outperform bigger competitors on return per ad dollar.

The weekly numbers

From the Daniks.AI weekly report, week of May 18–24, 2026, Amazon US:

Metric	Robot (Daniks.AI)	Human (Manual)	Blended
Impressions	115,940	945	116,885
Clicks	1,157	15	1,172
CTR	1.0%	1.6%	1.0%
Spend	$446	$10	$455
CPC	$0.39	$0.64	$0.39
Orders	149	1	150
Sales	$3,462	$20	$3,482
CVR	12.9%	6.7%	12.8%
ACoS	12.9%	48.3%	13.1%
ROAS	7.77x	2.07x	7.64x

And the number that ties the PPC program to the full business:

Metric	Value
Total ad sales	$3,482
Total revenue (organic + ad)	$5,639
TACoS	8.1%

Three things to notice. First, the robot side is doing essentially all the work — 99% of spend, 99% of orders, 99% of revenue. The manual track is not a competitor; it is a reference point. Second, every single efficiency metric on the robot side is exceptional: CPC under $0.40, CVR above 12%, ACoS under 13%, ROAS above 7.5x. Third, the TACoS of 8.1% means the brand is not PPC-dependent — organic revenue ($2,157) nearly matches ad-attributed revenue, and the total ad tax is only eight cents on the dollar.

What 12.9% ACoS at 7.77x ROAS actually means for the business

A 12.9% ACoS on an insulated drinkware brand — after Amazon’s 15% referral fee, FBA fulfillment fees on a mid-size product, and manufacturing costs — means the PPC program is deeply profitable on every order. The brand is not just covering customer acquisition cost; it is generating real margin on the ad-attributed sales themselves.

At 7.77x ROAS, every dollar of ad spend returns $7.77 in revenue. For a mid-range drinkware brand with roughly 35–40% gross margin before advertising, that means the ad spend itself is generating margin — not just paying for itself. The cost per order is $2.99 ($446 / 149 orders). On a $23 average order value, that $2.99 acquisition cost leaves substantial room for profit after all Amazon fees.

Compare that to the manual side’s economics: $10 spent on a single $20 order. The acquisition cost is $10 — nearly half the sale price. After Amazon’s fee stack, that order likely lost money. One order is not statistically meaningful, but the direction is clear: the manual side’s bid math is not working at this scale.

CPC: $0.39 in a competitive category

The robot side paid $0.39 per click. The manual side paid $0.64. That is a 39% gap — but more importantly, $0.39 CPC in the insulated drinkware category on Amazon US is well below the category average.

How does the AI keep CPC this low? The same mechanism documented across all Daniks.AI case studies, but amplified by maturity:

Months of keyword data. By the time a weekly report looks like this, the AI has been harvesting and pruning keywords for weeks or months. The surviving keyword set is lean — high-converting, low-competition long-tail terms that clear at lower auction prices.
Continuous bid adjustment. The AI does not wait for a weekly review to catch an overbid. It adjusts on a near-continuous loop, tracking the auction’s clearing price and never paying more than the conversion data justifies.
Aggressive negative matching. Non-converting search terms get negated within hours, not days. The spend that would have drifted toward those terms is redirected to the converting set.

The result: 1,157 clicks at $0.39 each, converting at 12.9%, producing $3,462 in sales. The efficiency compounds: cheap clicks × high conversion = very low ACoS.

CVR: 12.9% is where keyword maturity shows

One in eight clicks on the robot-managed campaigns turned into an order. That is an exceptional conversion rate for Amazon PPC in any category, and it reflects not just listing quality but keyword quality.

By the time the AI’s keyword inventory has matured — after weeks of the harvest-and-prune loop — the search terms driving clicks are overwhelmingly high-intent. The buyer who clicks on an AI-managed ad for NordFrost is arriving from a search term that the AI has already validated through conversion data. Broad, exploratory terms that generate clicks but not orders have been negative-matched out of the campaign set.

The manual side’s 6.7% CVR (1 order from 15 clicks) is not a meaningful data point — 15 clicks cannot produce a reliable conversion rate. But it is directionally consistent with what a manual campaign looks like without the same keyword maturity: fewer converting terms in the portfolio, more clicks from unqualified traffic.

TACoS: the metric that tells the real story

The previous Daniks.AI case studies focused on ACoS and ROAS — the advertising efficiency metrics. This NordFrost report introduces a different lens: TACoS (Total Advertising Cost of Sale), which measures ad spend as a percentage of total revenue, not just ad-attributed revenue.

NordFrost’s TACoS: 8.1%

Total ad spend: $455
Total revenue (organic + ad-attributed): $5,639

That 8.1% means the brand spent eight cents in advertising for every dollar of total revenue. For context:

TACoS below 10% is the range where PPC supports organic growth without dominating the cost structure. The brand has pricing power, organic ranking momentum, and room to scale ad spend if needed.
TACoS of 10–20% is the scaling band — the brand is investing more aggressively in PPC to gain market share, which is fine if ROAS remains healthy.
TACoS above 20% is the danger zone — the brand is PPC-dependent, and if ad performance dips, the entire business feels it.

NordFrost at 8.1% TACoS with 7.77x ROAS is the ideal configuration: PPC is contributing meaningful sales ($3,482 of $5,639 total) while consuming a small share of total revenue. The brand’s organic revenue ($2,157 in this week) provides a stable base that does not depend on ad spend.

The PPC strategy guide I run alongside any automation tool covers the TACoS framework in more detail.

How NordFrost compares to the other case studies

This is the fifth public Daniks.AI case study on TheFBAGirl. NordFrost is different from the others — not a balanced comparison but a steady-state snapshot of near-full AI management:

Case Study	ROAS (AI)	ACoS (AI)	CPC (AI)	CVR (AI)	Type
Tropeza	3.1x	32.7%	$0.68	2.48%	Parallel comparison
Basecamp	2.4x	42.0%	$1.02	14.5%	AI vs agency
Fornel	5.8x	17.1%	$0.42	7.50%	Parallel comparison
StellaHive	6.2x	16.1%	$0.40	10.65%	Parallel comparison
NordFrost	7.77x	12.9%	$0.39	12.9%	Steady-state snapshot

The progression is instructive. NordFrost has the best numbers of any published case study — and that is not because the brand is special. It is because the AI has been running long enough for the keyword set to fully mature. The earlier case studies capture the AI in its first weeks or months. NordFrost captures what happens after the optimization loop has been running and compounding.

Honest caveats

More caveats here than on the parallel-comparison case studies, because the data structure is different:

1. One week of data. This is a single weekly report snapshot, not a multi-week study. Weekly numbers fluctuate. The 7.77x ROAS may be higher or lower in the next report depending on competition, seasonality, and Amazon’s auction dynamics.

2. The human side has no statistical weight. Fifteen clicks and one order cannot produce a meaningful conversion rate, ACoS, or ROAS. The manual numbers are directional at best. I include them because the weekly report includes them, not because they are statistically valid.

3. Attribution window effects. Amazon’s 7-day attribution window means some orders in this weekly report may have been clicked in the prior week. The numbers are not perfectly contained within the May 18–24 window.

4. TACoS depends on organic strength. The 8.1% TACoS reflects the brand’s organic ranking and review velocity, not just PPC efficiency. A brand with weaker organic presence would have higher TACoS at the same ad spend.

5. I built the tool. Same disclosure as every other case study. I am the founder of Daniks.AI and I publish this because the numbers are strong. Run your own test.

What this means if you are running Amazon PPC

NordFrost represents the end state that the other case studies are moving toward. The pattern is clear across five published case studies:

Weeks 1–3: the AI collects data, builds the keyword inventory, starts pruning non-converters. ACoS is volatile.
Weeks 3–6: the keyword set tightens, CPC drops, CVR rises. You see Tropeza and Fornel-shaped gaps against manual.
Months 2+: the AI’s keyword inventory is fully mature. CPC is at or below category average, CVR is high, ACoS drops into the low teens. This is where NordFrost sits — the steady state.

If your brand is currently in the “weeks 1–3” phase with an AI tool and the ACoS looks messy, give it time. The compound effect of daily keyword harvesting and bid optimization needs three to six weeks of clean conversion data before the numbers look like NordFrost’s. If your ACoS has not improved by week six, the tool is not working for your category.

For the full feature breakdown, the Daniks.AI review covers the pricing tiers, what works, and where the tool falls short.

Frequently asked questions

What does 7.77x ROAS mean in dollar terms?

Every $1 spent on advertising returned $7.77 in ad-attributed sales. In this weekly report, NordFrost spent $446 and generated $3,462 in sales — a net contribution of $3,016 in revenue from the ad investment. After Amazon’s fee stack and COGS, the brand kept meaningful profit on those ad-attributed orders.

What is TACoS and why does it matter more than ACoS?

TACoS (Total Advertising Cost of Sale) measures ad spend as a percentage of total revenue — organic plus ad-attributed. ACoS only measures spend vs ad-attributed sales. TACoS tells you how dependent your business is on advertising. NordFrost’s 8.1% TACoS means ads are a small, efficient lever, not a crutch.

Is 12.9% CVR realistic for Amazon PPC?

It is at the high end but realistic for a well-optimized listing in a specific product category with strong purchase intent. Insulated drinkware buyers often search for exactly what they want — “20oz stainless steel tumbler” — and buy on the first click. The AI’s mature keyword set amplifies this by only bidding on terms with proven conversion history.

Why is the human side so small in this report?

The brand chose to allocate nearly all ad budget to the AI side based on prior performance. The $10 manual spend is a deliberately minimal reference track, not a competitive campaign. It shows what happens when a few leftover manual keywords run without active optimization.

How long does it take to reach NordFrost-level performance?

In my experience, most brands see their AI-managed campaigns stabilize in three to six weeks. The first two weeks are data collection. By week three, keyword harvesting is producing results. By month two, the keyword inventory is mature enough for NordFrost-level ACoS — assuming the listings are well-optimized and the category supports it.

Can every brand reach 7.77x ROAS with Daniks.AI?

No. ROAS depends on product price, margin, category competition, listing quality, and conversion rate. NordFrost’s 7.77x reflects a well-optimized listing in a category with moderate CPC and strong buyer intent. A high-CPC category like supplements or electronics would produce lower absolute ROAS even with the same AI optimization. The relative improvement over manual management is the more transferable data point.

What to do this week

Pull your TACoS from Seller Central — go to Business Reports, take your total revenue for the last seven days, and divide your total ad spend by that number. If your TACoS is above 15% and your ACoS is above 25%, your PPC program is consuming too much of your revenue. The fix is not spending less — it is spending more efficiently: lower CPC through better bid timing, higher CVR through better keyword targeting, both of which an AI agent does on autopilot.

Run a two-week parallel test on a clean campaign set. Keep your existing campaigns. Add Daniks.AI or any AI PPC tool on a separate track. Compare daily TACoS, not just ACoS — the total-revenue view tells you whether the AI is actually improving your business, not just your ad metrics.

For the operator-level breakdown of TACoS, ACoS, and the metrics that actually drive Amazon profitability, subscribe to @AmazonFBAGirl on YouTube. Real dashboards, real weekly reports, no hype. Drop a comment with your category and current TACoS — I read every one.