When discounting works (and when it quietly hurts you)

Discounting feels like a free lever. Lower the price, more people buy, problem solved. The catch is that discounts have a memory. Every time you train a customer to expect a discount, you raise the bar for what it takes to get them to buy at full price, and the math gets worse over time even if it looks better this quarter.

The cases where discounting genuinely works are narrower than most stores think. The first is clearance, where the alternative is paying to store inventory that's no longer aligned with the season or the catalog. Discounting until clearance is gone is rational. The second is acquisition, where a one-time first-purchase discount converts visitors who'd otherwise leave forever. The third is a deliberate, calendar-anchored event — Black Friday, an anniversary sale — where the rarity of the discount is part of why it works.

The cases where discounting quietly hurts are more common. Routine "free shipping over $X" thresholds tend to anchor the customer's expectations downward without noticeably increasing average order value. Site-wide percent-off banners that are always running teach customers that the listed price is fictional. Newsletter discount codes that go out monthly retrain the most engaged segment of your audience to ignore non-discounted launches.

The honest test of whether a discount is helping is whether the lifetime value of customers acquired during the discount is comparable to those acquired at full price. In most stores it isn't — discount-acquired customers churn faster, return more often, and don't stack up to a healthy second purchase. If you measure only the first transaction, every discount looks like a win. If you measure the second and third, the picture changes.

A useful alternative to running a discount is running a story. "We made this with a small batch of unusual leather we sourced once" outperforms "20% off all leather goods" for about the same amount of work. The story creates urgency without anchoring the price down, and the customers it attracts are the ones who would have paid full anyway — they just needed a reason to act now.

For stores that already have a discount habit and want to taper it, the fix is gradual. Stop the lowest-value discounts first — the always-on banners and the small percent-offs that aren't moving the needle. Replace them with one or two larger, calendar-anchored events. Watch six months of data before judging the change. The first quarter often looks worse on a top-line basis; the rebuilt margin shows up in months four through six.

The other lever is bundles. Combining items at a small price advantage — but never advertised as a discount — captures most of the conversion lift of discounting without training the price-anchor down. Bundles also tend to attract a different kind of customer: someone solving a project, not a sale-shopper, and projects have higher repeat rates.

The piece we'd recommend on this dives into a small-cohort analysis from a single store that ran discounts heavily for two years, then weaned off. The honest accounting of what they got back, what they didn't, and what they'd do differently is the kind of write-up that's harder to find than it should be.