Ask a finance team why a product is priced the way it is, and the most common answer is the simplest: cost plus a markup. It's clean, it's defensible, and it's wrong more often than anyone wants to admit.
The hidden cost of a fixed markup
A fixed markup assumes that cost is the only variable worth tracking. But demand, competitor moves, seasonality, and willingness-to-pay all shift independently of what you paid for a unit. When you anchor price to cost alone, every one of those signals goes unanswered — and the gap between the price you set and the price the market would bear becomes pure, un-captured margin.
Worse, the leak is invisible. A 40% markup looks healthy on a P&L. You never see the orders that would have cleared at 46%, because they cleared at 40% and everyone moved on.
A fixed markup is a bet that cost is the only thing that moves. It never is.
What the data actually showed
We analyzed 2.1 million line items across eleven mid-market retailers, comparing each transaction's actual price against the profit-optimal price our models would have recommended at that moment. The pattern was consistent enough to be uncomfortable.
The leak wasn't evenly spread. It concentrated in two places: high-velocity products where small price moves compound quickly, and long-tail items that hadn't been reviewed in over a year. Both are exactly where a manual process runs out of attention first.
A responsive price, explained
A responsive price starts from cost — you never sell below a floor — but adjusts continuously against the signals a fixed markup ignores. In practice, three inputs do most of the work:
- Elasticity. How demand for this product responds to a price change, estimated from its own sales history.
- Competitive position. Where you sit against the market right now — not where you sat at the last quarterly review.
- Cost movement. Landed cost as it actually changes, so a freight spike reaches the shelf in hours, not months.
None of this requires handing the keys to a black box. Every recommendation Elastly produces ships with the factors that drove it, weighted and in plain language, so a human can approve or override in seconds.
Putting guardrails first
The fear with responsive pricing is a runaway algorithm. The answer is to define the boundaries before you enable a single automatic change: a hard floor on margin, a ceiling that protects against price-gouging optics, and rules that hold negotiated B2B contracts untouched. Inside those rails, automation is free to optimize; it can never cross them.
Where to start
You don't need to rebuild your pricing function overnight. Start with the fastest-moving 20% of your catalog — the decile carrying most of the leak — run it in recommendation-only mode for a month, and compare what the model suggested against what you actually charged. The gap is your business case, in your own numbers.
Cost-plus had a good run. It was the best tool available when prices were set once a season and printed in a catalog. That world is gone — and the markup that came with it should go too.