Price Sensitivity and Promotional Effectiveness Research in South African Retail

Understanding how price changes and promotions drive purchase behavior is the difference between eroding margin and unlocking profitable growth. At Research Bureau we design and deliver rigorous, actionable price sensitivity and promotional effectiveness research tailored to South African retail—across grocery, FMCG, general merchandise, e-commerce and informal channels. Our work combines shopper insight, transaction analytics, and advanced econometrics to translate complex data into clear pricing and promotion decisions.

Why price sensitivity and promo effectiveness matter now

South Africa’s retail environment is highly dynamic. Consumers face rising living costs, VAT at 15%, and uneven income distribution that makes price a primary purchase driver for many households. Meanwhile, retailers and brands compete across dense promotional calendars and multiple channels—from national supermarkets to township spazas and online marketplaces.

Price and promo decisions influence:

Short-term sales velocity and margin realization.
Longer-term brand equity and perceptual price positioning.
SKU-level cannibalization and cross-product halo effects.
Inventory turnover, shrinkage, and supply chain efficiency.

Getting price and promotion strategy right requires rigorous measurement, confident causal inference, and actionable recommendations that align with local shopper realities. That is what our team delivers.

What we deliver — outcomes that drive profit

We translate research into decisions your commercial teams can execute. Typical deliverables include:

Price elasticity matrices by brand, SKU, channel and shopper segment.
Promotion lift models showing incremental sales vs. baseline demand and ROI.
Optimized price architecture that preserves brand positioning and margin.
Trade promotion optimization (TPO) plans and simulated promotion scenarios.
Dashboards & decision tools for pricing, category and trade teams.
Implementation roadmaps to integrate research outputs with ERP, pricing engines or promotion management systems.

Each deliverable is tailored to the South African retail landscape and validated using both experimental and observational data.

Our expertise and approach

We combine field research, commercial experience and advanced analytics. Our multi-disciplinary teams include pricing economists, data scientists, category strategists and in-market researchers who understand how South African shoppers behave in formal and informal retail settings.

Our approach is:

Evidence-first: we use experiments where possible and robust econometric controls where experiments aren’t feasible.
Shopper-centred: we link price sensitivity to real shopper segments, not just aggregate demand curves.
Action-oriented: every insight is paired with execution-ready recommendations, testing plans and KPI targets.

We adhere to strict data governance, privacy best-practices and transparent methodology so you can trust the outputs.

Research methodologies — when to use what

Below is a concise comparison of common pricing and promotional research methods and when each is appropriate.

Method	Best for	Key outputs	Time & cost
Van Westendorp Price Sensitivity Meter	New product price range / perceptual thresholds	Acceptable price range, optimal price point (POD)	Quick, low cost
Gabor-Granger	Price choice elasticity for single SKU	Demand curve, willingness-to-pay distribution	Moderate
Conjoint (Choice-based)	Bundles, multi-attribute pricing, SKU interactions	Simulated market share across price sets	Moderate-high
Sales Elasticity (Econometrics)	Real-world price promotions using POS data	Price elasticities, cross-price effects	Uses historical data
Promo Lift and Uplift Modeling	Measure incremental sales of promotions	Incremental units, ROI, cannibalization	Moderate-high
A/B / Store-level experiments	Causal inference on price/promo changes	Clean uplift and behavioral change estimates	Higher cost, longer timeline
Qualitative Shopper Research	Understand motivations behind price sensitivity	Shopper personas, triggers, messaging	Low-moderate

Detailed methodologies and when we apply them

Van Westendorp Price Sensitivity Meter (PWSP)

We use Van Westendorp for rapid perceptual pricing on new or rebranded SKUs. It identifies:

Too cheap, too expensive, cheap, and expensive thresholds.
A recommended indifference price point (IDP) and optimal price point (POD).

PWSP is best when perceptions matter—premium vs value positioning—and when you need fast directional insight across LSM segments.

Gabor-Granger pricing

Gabor-Granger experiments measure direct willingness-to-pay by presenting respondents with a series of prices and asking purchase likelihood. We segment responses to produce:

SKU-level demand curves.
Price elasticities by segment and channel.
Price sensitivity maps across metropolitan and township shoppers.

This method is ideal for iterative price testing before a market launch or for price re-sets.

Choice-Based Conjoint (CBC)

Conjoint analysis helps when price interacts with other attributes—size, pack format, bundle offers or promotions. We design CBC studies to simulate realistic trade-offs and to estimate:

Attribute utility scores.
Market share under alternative price/pack configurations.
Elasticities for bundles vs single-items.

Use conjoint for pack rationalization, multipack pricing, and developing cross-category promotional bundles.

Sales elasticity and econometric models

When we have transactional POS data, we build elasticity models using:

Time-series regression with controls for seasonality, holiday effects, and competitor activity.
Panel data and fixed effects for store-level heterogeneity.
Bayesian hierarchical models where data sparsity exists.

These models quantify how a 1% price change affects volume and revenue across SKUs, stores, and regions. They also reveal cross-price elasticities and substitution patterns important for cannibalization analysis.

Promotion effectiveness and uplift modeling

Understanding true incremental sales from promotions is critical to avoid disguised cannibalization. We deploy:

Control-test quasi-experimental designs using matched stores or time periods.
Propensity-score matching and synthetic control when randomized tests are infeasible.
Uplift modeling to identify which shoppers are incremental vs. coupon redeemers.

Outputs include incremental units, margin impact, and ROI for different promo mechanics (price-off, multi-buy, FOC, coupons).

A/B and randomized experiments

For high-impact pricing moves (e.g., national price cuts, new loyalty pricing), we recommend randomized experiments:

Store or online user-level randomization for causal attribution.
Pre-registered KPIs and minimum detectable effect calculations.
Pre- and post-analysis to test persistence of effects.

Experiments give the cleanest causal evidence and reduce implementation risk.

Qualitative fieldwork and shopper ethnography

Numbers tell you the “what”, but qualitative research tells you the “why.” Our fieldwork includes:

In-store observations and shopper intercepts in urban and township formats.
In-home usage and purchase-decision diaries for durable goods.
Focus groups to explore price perceptions and promotional recall.

This insight is used to build realistic conjoint attributes, interpret elasticity heterogeneity, and craft promo messaging that increases conversion.

Panel and behavioral data integration

We fuse traditional research with:

Retailer loyalty panels.
Household purchase panels.
Web and app behavioral logs for e-commerce conversion analysis.

This multi-source approach improves model accuracy and helps separate short-term promo spikes from long-term behavior change.

Promotional mechanics we measure and optimize

Promotions vary in type and impact. We measure and optimize across common mechanics:

Price-off / percent discount: measure elasticity and margin leak.
Multi-buy (e.g., 2-for-1, bulk promotions): identify cannibalization and net incremental volume.
Temporary pack sizes / temporary sticks: test unit price vs perceived value.
Display & feature placement: measure traffic-driven uplift.
Coupons & digital offers: estimate redemption lift and incremental basket size.
Loyalty pricing: segment elasticity by loyalty tier and estimate program ROI.

For each mechanic we measure incremental sales, margin impact, halo/cannibalization, and post-promo hangover effects.

SKU, basket and channel-level analysis

A profitable promotion strategy must consider SKU interactions and channel differences. We analyze:

SKU-level cannibalization and cross-elasticities to prevent margin erosion.
Basket-level effects to capture trip drivers and add-on purchases.
Channel-specific price sensitivity for supermarkets, discounters, e-commerce, and spazas.

This ensures that a promotion that drives sales in one channel doesn’t destroy profitability across the category.

Pricing architecture and consumer segmentation

A one-price-fits-all approach is risky in South Africa. We help design price architecture that balances fairness, profitability, and brand positioning:

Tiered pricing by pack size, geography, and channel.
Value packs for price-sensitive segments and premium lines for less price-sensitive shoppers.
Promotional cadence—optimal frequency and depth of promotions to maximize lifetime value.

Segmentation is anchored in observable metrics: LSM groups, frequency of purchase, basket spend, and shopping mission. We provide personas with recommended price and promo levers.

Trade Promotion Optimization (TPO)

Trade spend is often the largest line item for CPG brands. Our TPO services include:

Attribution of incremental volume to trade spend across accounts and channels.
Simulation of alternative promotion plans and spend reallocation.
Optimized calendar with promotion depth, timing and SKU selection to maximize net revenue.

We use constrained optimization to respect supply limits, account rules, and merchandising calendars.

Advanced analytics & machine learning for price and promo

We leverage advanced methods when data richness allows:

Bayesian hierarchical models for stable elasticity estimates across sparse SKUs.
Time-varying coefficient models to capture shifting price sensitivity during high inflation periods.
Uplift and causal forest models to identify which shopper groups are truly incremental.
Reinforcement learning simulations for continuous dynamic pricing and promo rules in e-commerce.

These models are built with interpretability in mind so commercial users can act on the outputs without requiring a data scientist on every decision.

Implementation & integration

Research is only valuable if integrated into commercial processes. We provide:

Export-ready elasticity tables and promo rules for pricing engines.
API-friendly dashboards and CSV exports for ERP and trade planning tools.
Training workshops for category, trade and pricing teams to interpret and operationalize insights.

We can also run pilot implementations, monitor outcomes, and iterate on the model post-deployment.

Reporting, KPIs and dashboards

Our standard KPI set includes:

Incremental units and incremental revenue from promotions.
Promo ROI: incremental margin / incremental trade spend.
Price elasticity by SKU, channel and segment.
Cannibalization rates and halo lift percentages.
Promo persistence and post-promo hangover.

Deliverables typically include interactive dashboards (Power BI, Tableau or web apps), and static executive summaries for board or buyer presentations.

Typical project types and sample timelines

We design engagements to fit your need and budget. Sample project types:

Rapid perceptual pricing (Van Westendorp): 3–4 weeks.
SKU price sensitivity (Gabor-Granger): 4–6 weeks.
Conjoint for bundles or pack redesign: 6–10 weeks.
Promo effectiveness with POS data (econometrics): 6–12 weeks.
Full TPO and implementation (end-to-end): 12–20 weeks with pilot phases.

Each engagement includes scoping, data ingestion, model build, validation, recommendations and handover.

Example case studies (anonymized)

Example 1 — FMCG brand: recovering margin through optimized promo depth

Problem: Heavy promotional frequency eroded margins despite rising volumes.
Approach: POS econometric model + store-level quasi-experiment.
Outcome: Identified 3 SKUs with >50% of promo sales cannibalized from base SKUs. Redesigned promo calendar and reduced depth by 5 percentage points, delivering a 7% margin lift with no net sales loss.

Example 2 — Retail banner: price architecture for township and urban stores

Problem: National pricing strategy produced underperformance in township stores.
Approach: Shopper segmentation, Gabor-Granger per LSM and basket analysis.
Outcome: Implemented targeted pack sizes and localised promos leading to 12% sales growth in targeted stores and improved availability with optimized replenishment.

Example 3 — E-commerce grocery: optimizing product bundles

Problem: Low conversion on bundles promoted during peak periods.
Approach: Choice-based conjoint and A/B tests on homepage placement.
Outcome: Identified preferred price/pack configuration and presentation; bundles converted 18% better and increased basket value by 9%.

Pricing and engagement models

We offer flexible commercial models depending on project scope:

Fixed-fee research projects with phased milestones.
Retainer for ongoing analytics and quarterly TPO management.
Outcome-linked engagements where a portion of fees ties to realized uplift or margin improvements.

Share your project details and we’ll provide a customised proposal and indicative budget.

Common KPIs we optimize

KPI	What it measures	Why it matters
Price elasticity	% change in volume per 1% price change	Guides price setting and revenue forecasting
Promo incremental units	Units sold above baseline	Separates true increment from cannibalization
Promo ROI	Incremental margin / promo cost	Determines profitability of promotional tactics
Cannibalization rate	Share of promo sales taken from other SKUs	Prevents internal competition and margin erosion
Basket lift	Change in average basket value	Measures cross-sell effectiveness during promos
Customer retention post-promo	Rate of repeat purchase after promo	Ensures short-term promos build long-term value

How we validate results and ensure robustness

We validate with multiple approaches:

Holdout tests and cross-validation for predictive models.
Replication across regions and channels to confirm stability.
Sensitivity analysis to test assumptions under different market scenarios.
Stakeholder walkthroughs and technical documentation for transparency.

Our models are designed to be conservative in uplift claims and to capture uncertainty explicitly.

Practical examples and calculations

Example — estimating promo ROI (simplified)

Baseline weekly sales for SKU A: 1,000 units at R20 = R20,000 revenue.
Promo: 20% off to R16, observed sales during promo week = 1,400 units.
Incremental units = 400 units. Incremental revenue = 400 * R16 = R6,400.
Incremental margin (assume gross margin 40% pre-promo, promo reduces net margin to 30%): incremental gross margin = 400 * R16 * 30% = R1,920.
If trade spend (cost of promo) = R1,200 in merchandising and display, Promo ROI = R1,920 / R1,200 = 1.6x.

This simplified example highlights how uplift can be profitable even when per-unit margin reduces, but the full analysis must account for cannibalization and post-promo effects.

Example — cross-price elasticity interpretation

If cross-price elasticity between SKU A and SKU B = -0.3, then a 10% price cut on A increases B’s sales by 3% (negative sign implies substitution effect). This helps determine whether promotions are expanding category demand or merely shifting share.

Frequently asked questions

How much data do you need for elasticity models?
- We can produce meaningful estimates with 6–12 months of POS data for fast-moving SKUs. For slower-moving items, we combine panel and experimental data or use hierarchical Bayesian pooling to borrow strength across similar SKUs.
Can you run tests in township spaza shops?
- Yes. We have field teams experienced in informal retail channels and can design culturally appropriate intercepts and small-store experiments.
Do you integrate with retail partners and ERP systems?
- We produce exportable outputs for common systems and can support APIs or CSV integrations as part of implementation.
How do you handle multi-channel promotions?
- We model channels jointly where data permits, and use channel-specific controls when necessary to isolate effects.

Next steps — get a tailored quote

Share a brief outline of your needs and we’ll propose a tailored scope and budget. Helpful details include:

Which SKUs, categories or channels you want analysed.
Available data types (POS, loyalty, web analytics, survey).
Desired timeline and key stakeholders.

Use the contact form on this page to request a quote.
Click the WhatsApp icon for a rapid chat with our pricing team.
Email us at [email protected] with a short brief and data summary.

We review leads within one business day and typically schedule a scoping call within 3–5 days.

Why partner with Research Bureau

Local market expertise: Deep experience across South African formal and informal retail sectors.
End-to-end delivery: From fieldwork and experiments to advanced econometrics and implementation.
Commercial focus: We prioritise net margin and lifetime value, not just vanity metrics.
Transparent methods: Models and assumptions are fully documented and reproducible.
Actionable outputs: Decision rules, dashboards and training so teams can act quickly.

We pride ourselves on creating pragmatic, evidence-backed price and promo strategies that deliver measurable commercial outcomes.

Final note

Price and promotion decisions sit at the heart of retail profitability. When executed without robust evidence they can erode margin and damage brand equity. When guided by rigorous research, local market insight and operational alignment, pricing and promotions become a powerful lever for sustainable growth.

Share your project brief now and let Research Bureau build a data-led pricing and promotion roadmap tailored to South Africa’s unique retail landscape. Contact us via the form, click the WhatsApp icon, or email [email protected].