PPC Reporting — Numbers That Decide

Part 1

The reporting mindset

Ask what the single most impactful PPC skill is and most people say bidding. The close second — and the one that separates operators from rockstars — is reporting and analytics. Not reading numbers; anyone can open the console and recite that ACOS rose five points. The skill is understanding the context behind the numbers, telling the story the trend is telling, and turning that story into action items that move the business.

The three-step framework

The five golden metrics

Always start with the raw ingredients. Jumping straight to calculated metrics on aggregated data is how misreads happen; the five raw numbers can't lie to you the way a blended average can.

The growth decomposition formula

When someone asks “why are sales up (or down)?” — decompose. In Amazon terms: sessions × unit session percentage × average selling price. Only three levers move sales; everything else rolls up into one of them. The analyst's job is to identify which lever moved and why.

Part 2

Critically thinking through metrics

CTR — the metric of first impressions

CTR measures how well your main image, price, reviews, and title win attention in results. It moves for five main reasons: placement mix (product-page impressions naturally carry lower CTR than top of search — a mix shift alone moves the average), main image quality (the biggest controllable factor — test it relentlessly), price competitiveness against competitor deals, new competition with flashier listings, and keyword breadth (broader terms mean less specific intent and lower CTR).

Conversion rate — the closer

CVR determines whether traffic becomes revenue. It moves with the funnel stage of your traffic (bottom-funnel terms convert several times better than browsing terms — mix shifts move the average), brand vs. non-brand mix (branded converts dramatically higher), the product's own strength (price, reviews, images, A+), competitor actions, and seasonality or macro conditions.

CPC — the cost of doing business

Your bid is only one input. The market — what competitors will pay — sets the price around you, and Amazon's relevance judgment acts as a quality score: a poorly matched product pays a premium for the same click.

ACOS and ROAS — outputs, not inputs

Part 3

The Sponsored Products reports

Amazon's console reports each slice the same underlying spend differently. Knowing each report's quirks — and its lookback limit — is half the craft.

Search term report

The queries that actually triggered your ads, against the keywords you targeted. The foundation of harvesting and negation. Quirks worth knowing: data reaches back roughly 65 days (verified — not the 60 everyone repeats), attribution is labeled 7-day for Sponsored Products, and one search term can appear across dozens of ad groups.

Search term impression share report

The search term report plus two extra columns: impression share — the percentage of available impressions your brand captured on that query — and impression rank — where you stack against every competing brand on the term. Together they turn the report from performance history into competitive positioning.

Sort by total possible impressions to find where volume lives, then cross-reference ACOS and CVR: high volume + low ACOS + low rank → push harder, profitably; high volume + high ACOS + low rank → fix conversion first, then push; rank #1 with high share → little room left, look elsewhere.

Advertised product report

Performance by individual ASIN, reaching back 90 days. Its best uses: per-product P&L (pair ad spend with Business Report sales for per-ASIN TACOS), share-of-total analysis (each product's slice of spend vs. its slice of sales — mismatches reveal over- and under-investment), same-SKU ratio (below ~50% and the product is mostly cross-selling, not building its own velocity), and CVR troubleshooting (sort ascending; the bottom of the list deserves investigation).

Campaign performance report

The most flexible report in the suite — daily, weekly, monthly, quarterly, or summary views reaching back years (hourly goes back 30 days). Combined with disciplined campaign naming, it becomes the most powerful tool you own: text functions can strip the ASIN, tactic, and target out of campaign names, letting you pivot years of history by any variable in the name — per-product trend reports no other report can produce.

Part 4

Search Query Performance — the best data on Amazon

If one data source changed how Amazon businesses get analyzed, it's SQP — the only place Amazon reveals organic-level, keyword-level data about your brand versus the entire market.

What makes it special

You'll find it in Seller Central → Brands → Brand Analytics (brand-registered sellers only). Monthly granularity is the sweet spot — weekly creates heavy exports for little extra insight.

Building the dashboard

The raw export needs work before it's useful. Download 12–18 months of brand-view data; add classification columns tagging each query brand / non-brand / competitor and by product category (a one-time effort — future months mostly reuse the same tags via lookup); then build pivots and charts for volume trends and share trends, filterable by tactic and category. What falls out is the closest thing to a market dashboard Amazon sellers can get.

The shipping-speed insight

One reconciliation note: SQP impression share will not match the ad console's impression share report, and that's expected — SQP leans toward organic impressions on page-one search, while the console counts all ad placements including product pages. They answer different questions; use each for its own question.

Part 5

Spreadsheet mastery for PPC reporting

At the core of PPC reporting is the spreadsheet — no macros or scripting required, but seven skills are non-negotiable:

Part 6

Data storytelling

The What → Why → So What framework

What: state the significant metric change — only the meaningful ones, no scorecard recitation. Why: drill to the root cause — never “due to optimizations.” So what: the concrete recommendation. That third element is the difference between an email people read and one they unsubscribe from.

Ask why seven times

Classic root-cause practice says five whys; PPC usually needs seven. Sales dropped 30% — why? Conversion fell 60% — why? Price moved from $15 to $22 — why? The brand raised MSRP on supply costs. Now the so-what: at this CPC we can't afford the current conversion rate; test a middle price that recovers enough conversion to hold rank while keeping the margin gain. The audience doesn't need to see all seven layers — they need to know you did the work and hear the conclusion.

Percentage of total — always

Allocation mismatches are invisible until you compute shares on both sides. Roll up to parent ASINs before judging — child-level shares mislead the same way child TACOS does.

Year-over-year context is non-negotiable

Never report month-over-month without year-over-year. December to January always looks like a disaster MoM; up 40% YoY with a smaller seasonal dip is actually outperformance. And check whether category search volume itself moved — if the market contracted 20% and you held share, that's a success story SQP can prove, not a failure to explain away.

Part 7

Building your reporting dashboard

Three essential layers

Why not just the ad console? It can't combine ad data with organic data; it can't show TACOS; it can't trend an individual ASIN over custom ranges or show instant YoY for any period. Every serious analysis ends up exported to a spreadsheet anyway — the dashboard just makes that permanent and always-on.

Tagging is everything

The most important reporting practice isn't the dashboard — it's tagging the account properly. Product tags (category, ASIN, line), tactic tags (brand, non-brand, competitor, auto), initiative tags (launch, seasonal push, ranking campaign). With clean tags, any view is one filter away; without them, every question is a fresh spreadsheet build.

Part 8

Client reporting that retains

Know your audience

A marketing lead wants keyword-level and campaign detail. A founder or CEO wants three bullets: are we growing, are we profitable, what should we do differently. The higher the audience, the more summarized and visual the report — bullets, not spreadsheets; simple charts, not 50-column tables; recommendations, not raw dumps. Jargon at the wrong altitude doesn't make you look smart; it makes you replaceable.

Always have a visual ready

Never walk into a performance conversation with only the ad console open. One prepared visual — chart, table, annotated screenshot — that answers the most likely big question changes the entire meeting. Fast sources: spreadsheet charts, an auto-refreshing dashboard, or annotated screenshots.

Reporting calls are sales calls

Every review is a retention event. Show the work, not just the results; preempt the concerns before they're raised; end with explicit next steps; follow up with a bulleted summary. You're demonstrating every time why the relationship continues.

When performance breaks

Things will go wrong; the analysis quality decides what happens next. The pattern that works: a top SKU collapses from 100 units a day to 5 — what happened is obvious, why is the job. Price jumped $15 to $22; conversion fell 60%; at the new conversion, current CPCs can't buy profitable visibility. The recommendation isn't “we'll optimize” — it's test $18.99: a middle point that may recover enough conversion to hold rank while keeping most of the margin. One chart showing price climbing as BSR falls makes the case undeniable.

The same discipline applies when the cause is the market: YoY sales down despite steady investment, SQP shows core-keyword volume down 25%, the category itself contracted. The recommendation shifts from growth mode to efficiency mode — trim the least incremental tactics, protect margin, watch monthly for recovery. Context turns a “failure” into a strategy.

The difference between a good PPC manager and a great one isn't bid-change volume or structural complexity — it's the ability to look at data and see what it means. Anyone can read a scorecard. The job is interpreter: raw data into strategy, context, and action. Data without a story is noise; a story without data is opinion. Together they're strategy — and if you want that kind of reporting on your account, the first diagnosis is free.