I made a mistake in a room full of people. The room was cold. The people were quiet. I was presenting a traffic pattern analysis for a mid-sized city. I had a chart on the wall. The chart showed the movement of cars. The chart showed that more cars were moving through the main street. I told the people that the movement was good. I told the people that the city was healthy.
Then the hiccups started. The hiccups were loud. The hiccups were rhythmic. I tried to talk through the hiccups. I tried to drink water. The hiccups continued. A woman in the front row asked a question. The woman asked where the cars were going. I did not have an answer.
I had counted the cars. I had not counted the destination. I had measured the activity. I had not measured the result. The hiccups made me look like a fool. The data made me a failure. I see this same mistake in engineering. I see the mistake on every invoice. I see the mistake in every board meeting.
The Disparity of the Screen
Priya sits in her office. The office is in a building for a logistics company. The company bought another company ago. Priya is the VP of Engineering. Priya has two screens. The left screen shows the September statement of work. The right screen shows the release notes.
The statement of work is from the vendor. The statement of work is long. The statement of work lists 682 hours. The statement of work mentions API modernization. The statement of work mentions discovery. The statement of work mentions iterative implementation.
The invoice visualization: 682 hours of “modernization” resulting in two endpoints.
Priya looks at the right screen. The release notes are short. The release notes mention two endpoints. The release notes mention one config change. The release notes do not mention the 682 hours. Priya looks at the August statement. Priya looks at the July statement. The statements are the same. The hours are always high. The shipping is always low.
Priya must go to an operating review on Tuesday. The board will be at the review. The CFO will be at the review. The CFO is a man named Mark. Mark likes the numbers. Mark will see the invoice for the 682 hours. Mark will ask what the company bought.
Priya tries to build a sentence for Mark. Priya wants to say the vendor is working hard. Priya wants to say the vendor is making progress. The sentence will not form. Every sentence Priya builds ends as a question. Priya cannot send a question to the board.
The Inverted Incentive
The vendor had a good quarter. The vendor billed for every hour. The vendor used the staff. The vendor paid the rent. The vendor made a profit. The vendor had a good quarter because nothing shipped. If the vendor had shipped the API in July, the vendor would have no hours to bill in September.
If the vendor was fast, the vendor would be poor. The vendor is not poor. The vendor is rational. People blame the vendor. People say the vendor is not competent. People say the vendor is lazy. This is not true. The vendor is competent. The vendor is following the contract.
The contract says the company will pay for hours. The company pays for the hours. The vendor provides the hours. The arrangement is performing correctly. The arrangement is a machine for turning time into money. The machine does not care about the software. The machine only cares about the clock.
I look at traffic patterns. I see cars stuck in a jam. The cars are not moving. The cars are burning fuel. The fuel costs money. If the driver of the car is paid by the hour, the driver does not mind the jam. The jam is a paycheck. The jam is a good quarter. The vendor is the driver. The software is the destination. The jam is the development process.
Priya scrolls through the tickets. There are 14 tickets closed in September. Four of the tickets are now open again. The vendor says the tickets needed more discovery. The vendor says the requirements changed. Re-opening a ticket creates more hours. Re-opening a ticket is a revenue event.
In a world of hourly billing, a bug is a future invoice. A mistake is a recurring revenue stream. We buy hours because hours are easy to count. A manager can look at a spreadsheet. The manager can see that 682 hours were worked. The manager can verify the hours.
The Comfort of the Discovery Phase
Ambiguity is comfortable. Ambiguity is a soft blanket. The vendor likes ambiguity because it allows for discovery. The buyer likes ambiguity because it hides the lack of a plan. The company spends millions of dollars to stay under the blanket. The company pays the vendor to stay in the discovery phase.
The discovery phase is the best phase for a vendor. The discovery phase has no release notes. The discovery phase has no bugs. The discovery phase only has hours. The CFO does not care about the discovery phase. Mark wants to see the return on the spend.
Mark calculates the cost of an endpoint. The cost is too high.
Mark asks Priya why the cost is high. Priya says the system is legacy. Priya says the code is brownfield. Priya says the code is difficult. These things are true. But these things are also excuses. The system is difficult, so the vendor takes more time. The vendor takes more time, so the vendor gets more money. The system being difficult is the vendor’s best friend.
The Outcome-Based Alternative
There is a different way to work. Some people do not sell hours. Some people sell the result. Some people join the team and commit to the code. These people use tools to move faster. These people use AI to handle the legacy code. These people do not bill for the time they spend being stuck. They bill for the progress they make.
is an engineering partner that works this way. They do not sell engineering capacity from the outside. They put a delivery pod inside the client team.
The pod has a senior engineer. The pod has an AI delivery architect. They join the daily stand-up. They commit to the client repository. They ship through the client pipeline. They work on the brownfield code. They use a framework to measure the work. They take a baseline in the first week. They look at the DORA metrics. They look at the velocity.
They commit to a lift in productivity. The lift is fifteen percent. They promise this lift within . If the lift does not happen, the invoice does not happen. This is the opposite of the hourly model.
Guaranteed within 60 days-or the invoice doesn’t happen.
I think about my hiccups. I think about the board room. If I had promised the board an outcome, the hiccups would not have mattered. If I had told them how to reduce the traffic by ten percent, they would have waited for me to stop hiccuping. They would have valued the result. They did not value my activity.
Priya needs to change the unit of exchange. Priya needs to stop buying the hours of the vendor. Priya needs to buy the lift in the team. Priya needs to tell the vendor that the invoice depends on the release notes. This is a hard conversation. The vendor will not like this conversation.
“A promise is a tension. When a brand says limited 16 times, the thread loses its memory.”
– Sofia, Thread Tension Calibrator
Priya should listen to the vendor. Then Priya should look at the empty release notes. The complexity of the system is the reason for the partnership. If the system were easy, Priya would not need the vendor. The vendor is paid to solve the complexity, not to document it for six hundred hours.
From Passenger to Engine
The invoice is a record of the time the vendor spent failing to reach the destination. The company is past the acquisition. The integration is slow. The technical debt is high. The debt is a weight. The hourly vendor is a passenger on that weight. The passenger is comfortable.
When you buy a measured outcome, the vendor becomes the engine. The vendor wants to reach the destination. The vendor wants to ship the code. The vendor wants the fifteen percent lift. The vendor uses AI because AI is fast. In the hourly model, AI is a threat to the vendor. AI reduces the hours. AI reduces the revenue.
AI is a threat to revenue. Speed is a punishment.
AI is a tool for profit. Speed is the reward.
Priya prepares her notes for Mark. She decides to stop defending the 682 hours. She decides to admit that the hours did not buy a result. She decides to tell Mark that the contract is performing correctly, and that is why it must change. She will propose a new model.
Mark will like this. The board will like this. The vendor might leave. If the vendor leaves because they cannot guarantee a lift, the vendor was never a partner. The vendor was a meter. You do not need a meter. You need an engine.
I stopped my hiccups by holding my breath. I sat down. I waited. The room was still quiet. I told the board I would come back with a plan for the destination. I told them I would stop counting the cars. I felt better. The truth is better than a chart of activity.
Priya sends the email. She cancels the discovery meeting for Wednesday. She asks for a shipping schedule. She waits for the reply. The screen is still bright. The release notes are still short. But the conversation has changed. The unit of exchange has changed.
The best quarter is no longer the one where nothing ships. The best quarter is the one where the company moves forward. I look at the traffic patterns now. I do not look for movement. I look for the end of the journey.
I look for the empty road because the people have arrived where they wanted to be. That is the only number that matters. The rest is just noise. The rest is just hiccups.