Reference Price Decay: Why Heavy TPR Quietly Collapses Your Non-Promoted Baseline

Every shopper carries an internal fair price for the brands they buy. Behavioural anchoring research calls it the reference price. It is the price the shopper expects to pay, the number that the actual shelf price gets compared against, the threshold above which a pack feels expensive and below which it feels like a bargain.

The reference price is not the listed shelf price. It is the price the shopper most often observes at the point of purchase. So when a brand is on TPR for 50 percent of weeks at 25 percent off, the most-frequently observed price is not the shelf price. It is somewhere between the shelf and the deal, weighted toward the deal.

Over 12 to 24 months of that cadence, the shopper's internal fair price drifts down. The full shelf price starts to feel like a price rise. The non-promoted weeks see softer volume. The promo weeks still spike, so the event-level report still looks fine. The damage hides inside a slowly sinking baseline.

What the data looks like in heavy-TPR categories

Across mainstream FMCG categories with heavy promotional cadence (frozen pizza, soft drinks, biscuits, household cleaning), the pattern is consistent. A brand running more than 40 percent of weeks on TPR at 20 percent or deeper depth typically loses 15 to 25 percent of its non-promoted baseline volume over 18 to 24 months. The brand still sells. It just only sells on promo. The full shelf price has become a sticker shock.

The mechanism is not new. It is anchoring and adaptation-level theory applied to retail pricing. Repeated exposure to a stimulus (the deal price) shifts the perceptual anchor. After enough exposure, the previous anchor (the shelf price) feels extreme.

Why this is a TPO Lesson 3 problem first

Reference price decay shows up in three places in the curriculum. Pricing Lesson 3 covers the threshold and round-number cliff side of the same psychology. Pricing Lesson 7 covers the promotional pricing trap from a strategic-pricing angle. But the working-team consequence sits in TPO Lesson 3, because Lesson 3 is where the baseline is measured. The baseline is the denominator on every TPR ROI calculation. If the denominator is decaying, every event ROI is being compared against a moving floor, and the post-event report cannot tell you whether last year was good or just less bad than this year.

The math behind reference price decay is not a single equation. It is a cadence loop where each event's depth and duration shifts the shopper's reference point a small amount, and the next event lifts from a slightly lower baseline.

A useful working approximation:

Baseline_t+1 = Baseline_t × (1 − k × promo_share × avg_depth)

Where:

• Baseline_t is the non-promoted weekly volume at the start of the period
• promo_share is the fraction of weeks running on TPR (0.50 for 50 percent of weeks)
• avg_depth is the average price-off across promo weeks (0.25 for 25 percent off)
• k is the category sensitivity coefficient, typically 0.04 to 0.08 for mainstream FMCG

The coefficient is empirical. It captures how reactive the category's shoppers are to repeated discounting. Premium-skewed categories (specialty coffee, premium chocolate) have lower k because shoppers anchor more on quality cues. Commodity-skewed categories (basic biscuits, household cleaning) have higher k because price is the dominant cue.

A 24-month worked example

Take a frozen pizza brand at 100 baseline units per week. Run 50 percent of weeks on TPR at 25 percent off. Use k = 0.06 for the category.

• Month 0 baseline: 100 units per week
• Month 6 baseline: 100 × (1 − 0.06 × 0.50 × 0.25)^6 ≈ 95.6 units (-4 percent)
• Month 12 baseline: ≈ 91.4 units (-9 percent)
• Month 18 baseline: ≈ 87.4 units (-13 percent)
• Month 24 baseline: ≈ 83.6 units (-16 percent)

That is a 16 percent erosion at the lower end of the typical range. Push promo_share to 0.60 with the same depth and the 24-month erosion runs closer to 20 percent. Push depth from 25 to 30 percent and you cross the 25 percent erosion line.

What the chart is telling you

The two lines start at the same baseline. The light-cadence brand (20 percent of weeks on promo) holds its baseline almost flat over two years. The heavy-cadence brand (50 percent of weeks on promo) loses about a fifth of its baseline volume across the same window.

The reported lift on each TPR event stays steady across both brands. Each event still shows a 2x to 2.5x spike against the most recent baseline. But the most recent baseline for the heavy-cadence brand is shrinking. So the absolute volume per event is falling. So the trade dollar buys less and less incremental revenue each cycle. So the brand keeps spending more to maintain the same nominal sales. So the spend escalates while the baseline collapses underneath.

Imagine a national frozen pizza brand at $5.99 shelf price, mainstream tier, $2.20 cost of goods. The brand was running 52 percent of weeks on TPR at an average depth of 27 percent off, mostly to defend a circular slot at three large grocery accounts. The TPR scorecards looked acceptable. Each event delivered roughly a 2.3x lift against baseline with positive ROI in the +5 to +15 percent range.

What the team had not been tracking was non-promoted weekly volume.

Year-on-year comparison once they pulled the data

• Year minus 2 non-promoted weekly baseline: 11,400 cases per week
• Year minus 1 non-promoted weekly baseline: 10,200 cases per week (-11 percent)
• Current year non-promoted weekly baseline: 8,900 cases per week (-22 percent vs Year minus 2)

The baseline had eroded by roughly a fifth over two years while the team was reporting "stable" promo performance. The reason the lift number had stayed steady was that the lift was being measured against a falling baseline. Every event still looked like a 2.3x spike. The 2.3x was just on top of fewer and fewer underlying units.

What they did

The brand pulled promotional support for one of two main SKUs for 16 weeks across two retailers. They held the shelf price flat. They reallocated half the freed-up trade dollars to gondola-end display funding on the same SKU during the holiday window.

What happened

• Weeks 1 to 4: weekly volume on the holiday SKU dropped 18 percent versus the prior promoted average. Painful for the sales team.
• Weeks 5 to 12: non-promoted weekly volume began rising. By week 12 it had recovered to within 4 percent of the prior promoted-average level, while running at full shelf price.
• Weeks 13 to 16: non-promoted weekly volume stabilised about 8 percent above the pre-holiday non-promoted baseline.

The shopper's reference price had moved back up. The brand re-introduced TPR on the holiday SKU at 35 percent of weeks (down from 52 percent) and the new baseline held.

What it cost

The 16-week holiday cost the brand roughly $290,000 in lost short-term gross profit (volume drop in weeks 1 to 4 net of trade spend savings). The recovered baseline added an estimated $1.1M in annualised gross profit through reduced promo dependency and a stronger non-promoted base. Roughly 3.8x return on the discomfort.

Reference price decay does not show up in the standard event-level scorecard. You have to look for it deliberately, in three places.

Track non-promoted weekly baseline as a top-line KPI

Most TPO teams report promo lift, incrementality, ROI, and trade spend. Almost nobody reports non-promoted weekly volume as a standalone line. Add it. Track the rolling 13-week average of weekly volume on non-promoted weeks only. Plot it against the same metric 12 months earlier. A flat or rising line means the reference price is intact. A line dropping more than 5 percent year-on-year is a reference-price decay alarm.

Cap promotional cadence as a portfolio rule

Set a hard rule that no SKU in the portfolio runs on TPR for more than 40 percent of weeks in a rolling 12-month window. The 40 percent threshold is not magic. It is the empirical line above which most heavy-TPR categories show measurable baseline erosion. Some categories tolerate 50 percent without visible decay (private-label-heavy categories where shoppers were never anchored on the brand price). Some erode at 30 percent (premium categories where the price is the brand). Pick the threshold based on category data, then enforce it.

Run a controlled "promo holiday" once a year

Pick one or two SKUs in the portfolio and pull all promotional support for 12 weeks. Measure non-promoted baseline at week 0, week 6, and week 12. If the baseline rises during the holiday, the prior cadence was eroding it. If it stays flat or falls, the cadence was probably appropriate. The holiday is uncomfortable for the sales team because it costs short-term volume. But it is the only clean diagnostic for whether reference-price damage is happening.

Connect the diagnosis to the lever you actually have

Reference price decay is a slow problem and a working-team team usually cannot fix it inside one promo cycle. The realistic levers are:

• Reduce frequency, not depth. Cutting from 50 to 40 percent of weeks at 25 percent depth slows decay more than cutting from 25 to 20 percent depth at the same frequency. Frequency is the bigger driver of anchoring.
• Replace TPR with non-price activation. Display, feature, and second-placement support drive uplift without re-anchoring the shopper on a deal price. Shift trade dollars from price-off to display where you can.
• Stagger across SKUs. If the brand must promote heavily, rotate which SKU is on TPR each cycle so no single SKU crosses the 40 percent threshold.
• Pair with shelf-price holds. When you do reduce promo cadence, hold the shelf price steady. Raising the shelf price during a cadence reduction is the one move guaranteed to accelerate the damage you are trying to repair.