Psychological pricing: charm prices, anchoring, and decoys that work
The four psychological pricing techniques with the most replicated evidence — charm endings, anchoring, the decoy effect, and bundling — plus the myths to ignore.
Luska Research
Pricing & Yield Team
Psychological pricing is the use of price presentation — not just price level — to shift demand. Done well, it lifts conversion without touching your margin. Done badly, it makes your brand feel cheap. Here are the four techniques with the most replicated evidence behind them, and how to use each without insulting your customers' intelligence.
1. Charm pricing (.99 endings)
The classic. Pricing at €9.99 instead of €10.00 reliably lifts conversion in mass-market consumer categories. The mechanism is the "left-digit effect" — buyers anchor on the leftmost digit and process €9.99 as "nine-something", not "basically ten".
Field experiments by Eric Anderson and Duncan Simester in a women's clothing mail-order catalogue found that a $9 price ending increased demand in all three experiments, with a stronger effect on new items and a weaker one when a "Sale" cue was also shown.
When NOT to use it
2. Anchoring
Show a high reference price first. Every subsequent price is judged relative to it. The anchor doesn't need to be your price — it can be a "compare at" RRP, a competitor's price, or a higher-tier product in the same lineup.
SaaS pricing pages exploit this constantly: the €499/mo "Enterprise" column makes €99/mo "Pro" look reasonable, even though Pro is the plan they actually want you to buy.
Without an anchor, customers compare your price to nothing — which means they invent their own anchor, and they're usually wrong.
3. The decoy effect
Add a third option that's strictly worse than the option you actually want to sell, and watch sales of your target option rise. The canonical example is The Economist's subscription experiment:
- Web only — $59
- Print only — $125
- Print + Web — $125
In Dan Ariely's classroom test of this offer, described in Predictably Irrational and widely summarised, most students chose Print + Web when all three options were shown. With the "Print only" decoy removed, most took the cheap web-only tier instead. The decoy did nothing except make the bundle look obviously correct.
4. Bundle and partition pricing
Bundling hides the per-unit price by selling a package — useful when individual items have high price-comparison risk (any branded SKU on Amazon).
Partitioning does the opposite — splits a single price into components ("€39 + €4 shipping" vs "€43 shipped"). Partitioning generally lifts perceived value when each component is small relative to the headline; it backfires when add-ons feel like nickel-and-diming (every airline ancillary, ever).
What the evidence does NOT support
- Always end in 7. No, the alleged "magic of seven" is folklore. Use whatever your A/B tests support; in most consumer categories that's still .99.
- Removing the currency symbol. Cited from one Cornell restaurant study (2009), whose authors warned the result may apply only to lunch at that one restaurant. For most e-commerce, omitting the currency just confuses international visitors.
- "Pay what you want" defaults to revenue. Outside a handful of indie-music edge cases, it doesn't.
How to apply this without becoming gimmicky
- Pick one technique per page. Stacking three of these on a single PDP looks desperate.
- A/B test, don't decree. Charm pricing wins in most of e-commerce, loses in luxury. Your category is your category — test it.
- Pair with rounded MAP. If your suppliers set MAP, your charm price still has to respect it — see MAP pricing compliance.
- Round wisely on AI/dynamic outputs. An algorithm that outputs €37.42 should snap to €37.99 (or €36.99 if your floor allows). Numbers that look "computed" leak the algorithm and feel cold.
Apply psychological rounding to every dynamic price
Luska rounds its price recommendations to the rule you choose (.99, .95, .90, .50 or whole numbers) so dynamic prices still look human.
See AI pricing