WWorld Markets
MarkovPolicyBondsEquitiesOffersPricingAmazonvs H10/JSPharmaHealthcareInsuranceLeaderboardBayesianChatCasesQuizSellersBuyersJob Search

Keyword Opportunity Matrix

For every keyword: search volume × rank-CTR × win-probability × predicted conversion → expected orders, revenue, and profit at the elasticity-optimal price. Competition is a composite of KD, page-1 authority, review depth, and sponsored density, shrunk toward the category mean.

Est. Monthly Traffic
41,333
Est. Orders / Mo
507
Est. Revenue / Mo
$26,366
Est. Ad Cost / Mo
$47,717
Est. Profit / Mo
-$37,570
Filters & predictive controls
Market scenario overrides

Applied to every keyword before scoring. Use to stress-test rank slippage, ad-cost inflation, fee hikes, or a more elastic demand curve.

Baseline

+ pushes you deeper (lower CTR); − promotes you.

Auction-cost multiplier on every paid keyword.

Additive shift to the marketplace referral %.

Scales per-unit fulfillment cost.

>1 = more price-sensitive, thinner Lerner margin.

Keyword worksheet
PhraseCategoryIntentVolumeKDTop DACPCTarget rankPaid?Comp.Win PCTRConv.TrafficOrdersRevenueProfit$/1kOpp.
0.7228.8%5.00%4.22%4,88759$4,634-$4,956-$54.77
20
0.6343.1%7.00%2.29%3,01230$1,839-$2,039-$61.59
26
0.7921.8%3.50%5.55%5,77570$2,381-$9,493-$57.53
16
0.5257.2%10.00%1.89%2,44226$1,451$685$30.87
72
0.5747.5%7.00%2.46%3,26038$1,597-$2,969-$73.32
28
0.6734.2%5.00%6.39%3,88585$2,382-$5,903-$79.77
23
0.6936.5%7.00%4.62%5,08286$5,309$2,375$39.26
63
0.7721.6%3.50%4.99%3,92742$2,628-$11,391-$103.55
16
0.5159.5%10.00%2.38%2,71038$1,611$622$22.97
73
0.7424.9%3.50%2.05%6,35332$2,534-$4,502-$27.29
18

Comp. = 0.45·KD + 0.25·DA + 0.20·log(reviews) + 0.10·SponsoredDensity, shrunk toward the category mean. Win P = σ(2·(1 − Comp) + (yourDA − topDA)/50 − 0.2·(rank − 1)). Traffic = Vol · CTR@rank · seasonality. Profit = Traffic · WinP · Conv · AOV · marginAt(p*) − adCost (paid only).

Explainability — smart plug alexa
Electronics · transactional
Win probability drivers
z = 2·(1−Comp) + (yourDA−topDA)/50 − 0.2·(rank−1) → Win P = σ(z) = 28.8%
Competition headwind (z-base)
+0.55
Authority edge (yourDA − topDA)
−-0.66
Rank penalty (target #5)
−-0.80
Total log-odds z
−-0.91
James-Stein shrinkage on competition
Comp_shrunk = (1−k)·Comp_raw + k·CatMean, k = 0.30
KD contribution (0.45·KD/100)
0.279
Top-page DA (0.25·DA/100)
0.195
Review depth (0.20·log₁₀(1+R)/5)
0.185
Sponsored density (0.10·s)
0.060
Comp_raw
0.719
Category mean
0.732
Comp_shrunk
0.723
Traffic decomposition (multiplicative → log shares)
Traffic = Vol · CTR@rank · Seasonality = 4887/mo
Volume (90,500)
134%
CTR @ rank #5 (5.00%)
-35%
Seasonality (1.08×)
1%
Predicted conversion rate
Conv = clip( convBase · intentMult · (0.85 + 0.5·CPCsignal) ) = 4.22%
Category base (Electronics)
2.80%
Intent multiplier (transactional)
1.60×
CPC commercial signal (CPC=$1.35)
0.94×
Elasticity margin (Lerner rule)
p* = (c + f_ful) / (1 − f_ref) · η/(η−1); η=2.40, markup=1.71×
Optimal p*
$19.56
Unit cost
$6.50
Fulfillment fee
$3.20
Referral (15.0%)
$2.93
Gross / conv
$6.93
Margin
35.4%
Orders / mo
59.4
Ad cost / mo
$6,597
Kalman update on search volume
Prior = baseline volume; observation = your override. Posterior blends them by the Kalman gain K.
Prior (baseline)
90,500
Observation
90,500
Gain K
0.845
Posterior
90,500

σ_prior ≈ 31675, σ_obs ≈ 13575. K near 1 means we trust your input; K near 0 keeps the baseline. Traffic and every downstream metric use the raw input; the posterior is shown for reference so you can see how much of the change is signal vs noise.

What changed the prediction (vs baseline)
One-at-a-time counterfactual: rebuild profit with each input reset to its baseline value; the gap is that input's contribution.

No inputs have been changed from the baseline for this keyword.

Greedy paid-budget allocator
PhraseSpendProfitProfit / $ spend
portable neck fan$2,861-$2,039-0.71x
smart plug alexa$2,139-$1,607-0.75x

Allocates the monthly budget across paid keywords in descending order of profit / ad-dollar until spent.