Ask the Junkless Data
Rule-based engine that computes answers from the fitted models — not pattern matching against canned strings. Live LLM integration is the next step.
Brand share of $ sales
Total dollar sales across the active filter set
Junkless weekly $ sales
Weekly Junkless dollars (filtered)
Brand performance summary
Computed on the current filter
How it works: Per (SKU × retailer) we fit log(units) = α + βp·log(price) + βd·display_share + βf·feature_share + ε on 104 weeks of Nielsen data. The slider plugs your inputs into that retailer's fitted equation. Predictions are only valid inside the observed price range (shown). Outside that, the model extrapolates and uncertainty widens fast.
Junkless SKU
Retailer
Avg unit price
$3.99
Display share (% of units)
Feature share (% of units)
Predicted units / week
—
—
Predicted $ / week
—
—
vs. avg week
—
vs current avg
Predicted units and $ across the price range
Model predictions at every price within the observed range, holding display/feature at current setting
Revenue-maximizing price (within observed range)
Price that maximizes predicted $ given current display/feature inputs
Elasticity model: Per (SKU × retailer) OLS on log(units) ~ log(price) + display_share + feature_share. Only models with n ≥ 15 weeks and meaningful price variation are shown. ★ = significant at p < 0.05.
Own-price elasticity by retailer (6ct / 6.6oz)
More negative = more price-sensitive at that retailer
Display & feature lift coefficients
Change in log-units for a 1.0 increase in display or feature share
Junkless price–volume scatter
Each dot = SKU × retailer × week. Color = % units on promo.
All fitted models
Promo lift method: For each (SKU × retailer), compare avg units in weeks where price-decrease share < 10% (baseline) vs weeks where it's > 50% (heavy promo). Lift % = (promo − base) / base. 95% CI from standard error of the difference in means. Cells with < 3 baseline or promo weeks are excluded.
Promo lift by retailer (6ct)
% lift in weekly units, heavy-promo vs baseline
Discount taken vs lift achieved
X = % price discount, Y = % volume lift. Above the dashed line = lift exceeds discount.
Promo / Display / Feature share by brand
Junkless merchandising mix by retailer
Promo lift detail table
Cross-elasticity model: For each brand, log(units) regressed on log-price of all 5 brands plus retailer & month fixed effects. A positive cell at row=Junkless, col=Kodiak means: when Kodiak raises price, Junkless gains units (substitute). Negative cells = complement or shared promo cycle.
Cross-elasticity matrix
Row = brand whose units are being predicted · Column = brand whose price is the input · ★ = significant at p<0.05
Junkless interactions ranked
Junkless $ sales by retailer
Junkless avg price by retailer
Full retailer breakdown
Source & method: Qualitative themes synthesized from web search across Walmart/Amazon review pages, product blogs, FDA recall notices, class-action news, and trade publications. Limitation: not direct review-API ingest — themes may over-index on high-signal events (recalls, lawsuits). For quantified sentiment, wire to a Brandwatch / Sprinklr / Reddit-API feed.
Brand sentiment & risk landscape
Per-brand positives, negatives, and the single biggest risk overhang
Category trends
Macro signals from Snack & Bakery, Innova, Statista, and news coverage
Strategic implications for Junkless
Where the qualitative landscape meets the Nielsen data — actionable opportunities and risks
Data limitations & next-step upgrades
Note: Every claim below is derived from a regression model or hypothesis test, not from raw means. Sample sizes, confidence intervals, and significance markers are shown inline.