8.1 Marketing analysis

If price rises from $4.00 to$4.60, price changes by 15%. With PED = −1.4, forecast quantity change = −1.4 × 15% = −21%. If income falls 6% and YED = 1.5, forecast demand change = −9%. If promotion rises 20% and demand rises 8%, PrED = 8% ÷ 20% = 0.4.

PED and total revenue, other things equal: elastic demand means price and revenue tend to move in opposite directions; inelastic demand means they tend to move in the same direction; unit elasticity leaves revenue broadly unchanged. Revenue = price × quantity, but profit also depends on variable/fixed cost, capacity and promotional cost.

Limitations: historical/small or correlation-based data; percentage-base and measurement error; elasticity changes by segment, geography, time horizon, price/income/spend range and product life cycle; competitors, substitutes, complements, quality, distribution, brand and external shocks also change demand; response may lag; simultaneous marketing changes make attribution difficult.

Stage-gate process: identify need/objective → generate ideas → screen against strategy, customer value, ethics/law and capability → develop/test concept with target users → business analysis (demand, price, costs, cash/break-even/risk) → R&D/design/prototype and technical testing → test market/pilot and refine product/marketing mix/operations → decide scale/launch → commercialise, monitor feedback/performance and improve or withdraw.

Cross-functional alignment is essential: marketing defines need/segment/offer; operations tests quality/capacity/supply; finance tests funding/cash/return; HR secures skills/teams; legal checks safety/IP; R&D converts knowledge into designs. A technically successful prototype can still fail commercially or operationally.

Judge importance against competition/product life cycle and pace of change, customer need/price sensitivity, brand/IP, business objectives and risk appetite, finance/skills/time, operational scale/distribution and alternatives such as process improvement, marketing or acquisition. Use staged investment and stop/learn criteria rather than treating sunk cost as a reason to continue.

A time series may contain trend (long-run direction), seasonal variation (regular within-year pattern), cyclical movement (business-cycle pattern) and random variation. Four-period centred moving averages smooth quarterly seasonality to estimate trend-cycle; they do not explain its cause.

Four-period centred moving-average method: (1) add four consecutive quarterly observations and divide by 4; this moving average lies between the two middle quarters. (2) Move forward one quarter and repeat. (3) Average two adjacent four-period moving averages to centre the result on their shared middle quarter. Example: 4-MAs 120 and 128 give centred MA = (120 + 128) ÷ 2 = 124.

For an additive model: seasonal variation = actual sales − centred moving-average trend. Group variations by quarter and average each quarter across years (adjust rounding so four quarterly averages sum to zero if required). Forecast = extrapolated trend + average seasonal variation for that quarter. Example: trend 124 plus Q4 seasonal variation +18 gives forecast 142.

Use quantitative history as a baseline, then explicitly adjust/scenario-test known changes in price/promotion, products, competitors, capacity/distribution, economy/law/technology and one-off shocks. Compare forecast with actual, investigate error and update assumptions; use ranges or high/base/low cases for decisions with different downside risk.

Time-series limitations: past may not continue; few/inaccurate data and outliers; moving averages lose endpoint observations and lag turning points; different windows change smoothness; seasonal averages change; new products/markets lack history; external/competitor/marketing/capacity changes are omitted; extrapolation and false precision. Qualitative forecasts add current insight but bias and politics.