Who gets the capital when the new offer is still worse?

Blockbuster's revenue sat above $5 billion, and late fees alone brought in $800 million, when Netflix offered the chain a deal at $50 million. Blockbuster passed. Netflix at that point carried mounting losses and an uncertain future, and it could not hand a customer a disc the same afternoon. It could mail a wide library to people who lived outside the core store map. The Christensen Institute uses that pass to show a sequence incumbents keep misreading. A cheaper offer takes root in simple uses, usually because more people can reach it, and then it improves until it displaces the firm that owned the category.

How the cheaper offer moves up

The institute's account has a direction. The offer starts at the bottom, or with people the current product never served, and it moves upmarket. Along the way it makes the product or service more accessible and more affordable, so a larger population can buy it. A breakthrough that makes a good product better, for the customers who already pay the most, sits on a different path. The institute's example of that path is every new version of Apple's iPhone. Those versions keep the current contest going. They sell for more money, at a higher margin, to people who already wanted a better phone.

Integrated steel shows the upmarket climb with the margins attached. In the mid-1960s, mini-mills began melting scrap in electric arc furnaces. The process cut costs by 20 percent. At first they could make only the lowest grade, rebar. Integrated mills earned a 7 percent gross margin on rebar, so they let the mini-mills take it. The mini-mills earned gross margins above 20 percent on that same low-grade product. They then climbed into structural steel and into sheet. By the early 2000s, Nucor had turned the climb into a full displacement of integrated makers such as Bethlehem Steel and US Steel.

No integrated company managed to run mini-mill technology inside its own model, even with that cost advantage sitting on the table. The institute's reading is the one a CFO should sit with. At each step the mini-mill did not look like a threat. Rebar was a thin business for the integrated mill. The spreadsheet said the same thing every year: spend a little money and point capacity at higher-margin grades, and leave the furnaces to the newcomer. Both sides believed they had made a sound financial decision. Both were trying to maximise profit. The choice kept working until the integrated mills ran out of customers.

That climb, from a weak cheap offer to a full displacement, is the S curve an incumbent has to answer. The early offer looks worse on the measures the core business prints. Same-day rental from a store beat a disc in the post. A sheet from an integrated mill beat early rebar from a furnace. The offer then improves, while you keep funding the comparison you already win.

The filter inside the firm that already leads

A corporate innovation programme usually sits inside the firm that already has the high-margin customers. The people who release capital have a live P&L to defend. If the new offer really is on the disruptive path, it will miss the performance chart those customers use. It will cost less, or it will reach people the core does not serve, and it will look plain on the scorecard the core defends. The six checks on the institute's page are a way to see that shape before a lab asks for another year of funding.

1. Does the offer target nonconsumers, or people the current product already overserves?
2. Does it look worse than the current offer on the historical measures of performance?
3. Can a customer use it more simply, and does it cost less to reach?
4. Does a technology exist that can carry it upmarket as it improves?
5. Does a business model sit under that technology, so the unit can sustain itself?
6. Do today's providers have a reason to ignore it at the start, because it does not threaten the profit they came for?

If the offer fails those checks, you may still have a useful project. You have a sustaining project, and the core business is the right owner. The stall that wastes a programme is the other case. The offer matches the checks, and it still has to win capital in the same review that funds the next version of the current product.

That review has a rational script. The core customer wants a better version and will pay a higher margin for it. The bottom of the market pays less. Someone the firm does not serve yet pays nothing this quarter. A sponsor who brings a cheaper offer into that room is asking the firm to accept a lower margin on purpose. The room refuses. Or it agrees, and then it forces the offer to grow the cost structure of the core, until the offer resembles the product it set out to sit beside.

Blockbuster could not detach from the physical locations. The locations were expensive, and the model wanted to use them. The chain had more than 9,000 stores and 60,000 employees. A DVD-by-mail service that also had to feed that footprint could not run at Netflix's scale. The institute lists the model Netflix actually ran: mail order from a distribution centre, with no late fees and no stores. Mail and the internet enabled it. The postal service and direct sales on the internet carried the value network. The business model was the part Blockbuster would not copy, because copying it meant dropping the late fees and the stores.

Toyota's Corona is the same filter in a shorter form. A cheap, tiny car arrived in the 1960s. General Motors and Ford kept building big cars for wealthier buyers, and left the cheap car alone while the margin sat upmarket. Desktop computers later put enough power on one person's desk to thin out the reason for a cabinet-sized machine in each department.

What the programme keeps bundling together

The institute separates kinds of innovation so a firm can stop running them through one hurdle rate. Sustaining work improves the current solution for people who want better performance, and it often sells for more money. Efficiency work changes the process, keeps the same customers and the same model, and frees cash. Toyota's just-in-time system is their example of that. Low-end disruption enters an existing market at the bottom with a cheaper model, and the incumbent walks upmarket because the bottom pays too little to fight for. CVS MinuteClinic is the example they name. New-market disruption builds a segment outside the current market. Their example is the smartphone set against the laptop.

An innovation programme stalls when leaders ask one team to deliver this whole list through one set of customers and one margin floor. Sustaining work and efficiency work belong next to the P&L that already knows those customers. The disruptive offer needs a model that can live on the lower margin, or on customers the core does not serve, without borrowing the core's cost structure. If the new unit must keep the stores and the late fees, the spreadsheet will keep starving it, and the spreadsheet will look responsible while it does.

You can already have fixed an earlier stall and still hit this one. Why large organisations get better at process and worse at making something new is about the pattern where process people end up in charge and workshops stand in for a product. Get past that, give someone the authority to ship, and you can still lose the next curve, because the thing you ship has to beat the core margin in the same year. A proved pilot raises a later question. What is missing after a pilot has already worked asks who in the line will own the bet. The question here comes before that handoff. Will you let the offer stay small and cheap, on a model the core does not have to approve, for long enough that it can move up?

A lean startup and product sprint can force a decision on the riskiest assumption in a week. Point it at the business model. The assumption that matters is whether this offer can survive without the core's stores, fees or plant. A sprint that only tests a nicer version of the current product funds sustaining work and calls it a response to the shift.

What to take into the next capital review

Take the offer you are about to fund and read it against the sequence.

1. Name the customer the current product overserves, or the person the firm does not serve yet. If the only name on the slide is last year's key account, you are funding a sustaining upgrade, and the core team should own it.
2. Write the measure on which the new offer looks worse. If it already wins on the core scorecard, it has a different shape from the one the institute describes. Say so, and hand it to the core team.
3. Name one cost you will refuse to copy from the current model. A store network, a fee the core loves, a plant that must stay full or a margin floor taken from the best product. If you cannot name one, the new unit will inherit the old costs in the first planning round.
4. Say which technology will carry the offer upmarket, and what better means in a year on the new customer's terms.
5. Give the unit a profit rule it can meet at the bottom. A hurdle copied from the core will kill the offer in the first comparison, and the kill will look like discipline.

I would keep the core team on the sustaining work. They are good at it, and the customers who pay this year's margin still need the next version. I would put the cheaper offer in a unit that can post a small profit on the customers the core will not fight for. Netflix's later worth, on the institute's figures, sits above $197 billion, and Blockbuster filed for bankruptcy after it clung to the stores. Both firms could tell a profit story. Only the new model was free of the old footprint.

If your programme cannot pass a cheaper offer that looks worse on the current scorecard, and house it where the old cost structure cannot reach, the capital will keep flowing to the next version of what already sells. The review will look tight. The next curve will get built by someone who does not have your stores to protect.