This is the Electi review. Today, retailers keep running promotions the same tired way, and it's quietly eating their profit. Here's the core argument from Electy's new piece on promotion optimization. Most retail promotions are still planned on habit, gut feel, and calendar tradition rather than data. A category manager sees last October we did 20% off. So this October gets 20% off again, no measurement of what that discount actually did to volume, to margin, to cannibalization of full price sales. The article's position is blunt. That approach is a profit leak, not a growth lever. The fix it lays out has three parts. First, plan promotions using data before they launch, not after, modeling expected uplift instead of copying last year's calendar. Second, actually measure lift. Not just sales went up, but incremental lift. How much of that volume increase wouldn't have happened without the discount, and how much was cannibalized from adjacent products or pulled forward from future full-price purchases. Third, protect margin as a hard constraint, not an afterthought. Testing promotion depth and mechanics against margin impact, not just top line revenue. AI and testing are framed as tools to do this at scale, running structured tests across categories, comparing forecasted versus actual lift, and catching promotions that move volume but destroy profitability before they become permanent fixtures on the calendar. What's notable is the skepticism embedded in the piece itself. It treats promotion and profit as frequently opposed, not aligned by default. A promotion that grows sales but shrinks margin isn't a win by this framework. It's a miscalibrated cut. The whole discipline described is essentially retailers learning to distinguish a promotion that works from one that merely happened. No revolutionary technology claim here, the argument is about process discipline, plan with data, measure lift honestly, and let margin protection veto anything that fails the test. Simple to state. That's the review.