Aligned Software vs. Extraction Software

Overview

Call the two shapes aligned and extraction. It is a description of incentives, not a moral grade, and most real vendors sit somewhere between them rather than at either pole.

Aligned software earns more when you get more value from it. You add people, you pay for people. You use a bigger piece of it, you pay for the bigger piece. The base version is a real, working product rather than a demo with the useful parts removed. The price is published, so the number you pay is the number anyone pays. Your data is yours, in a format something else can read, available without asking. The vendor keeps you by continuing to be worth it.

Extraction software earns more when your alternatives get worse. The published price is for the first year. Features that once worked migrate up a tier. Your own records live in a format only that product opens, and the export is a support request rather than a button. The functional bill grows without the value growing, because the growth is coming from your reduced ability to leave rather than from anything new you received.

Neither shape is announced. Both are inferable, and the inference costs about an hour.

The Challenge

The problem is that every signal available at purchase time is produced by the party selling to you. The demo is theirs. The pricing page is theirs. The case studies are theirs. The one document that would actually settle it, the agreement governing what happens on renewal and on exit, is usually the last thing you see and the thing you are least equipped to read, arriving at the end of a process where saying no has become socially expensive.

So the model reveals itself later, on a schedule you do not control. It shows up as a renewal quote with no relationship to last year's. As a feature you were using appearing behind a tier you were not on. As a five-year archive of your own records that turns out to be readable only from inside a subscription you were about to cancel. None of these is a scandal. Each is a business model working exactly as designed, discovered at the one moment when your options are worst.

The fix is not skepticism, which mostly produces slower bad decisions. It is moving a few cheap questions from after the signature to before it.

Why It Matters

Four reasons this is worth an hour of a practice manager's time:

Lock-in is rarely a clause. It is usually a file format. People look for the trap in the contract, and the contract is often fine. The real constraint is that six years of records exist in a shape nothing else reads, so leaving means retyping history. That is not written anywhere. It is a product decision, and you can detect it in ten minutes by trying an export.

The published price and the price you will pay are different questions. A number on a pricing page describes acquisition. What matters over five years is what the renewal looks like, whether that renewal figure is also published, and whether it is the same for a customer who negotiates and one who does not. A vendor that publishes both has given up a lever on purpose, and that is worth noticing.

This is already your compliance homework. For a regulated practice, vendor oversight is not optional, and every business associate touching your records is supposed to be evaluated and documented. The questions in this piece are the same questions that oversight asks, so running them is not extra work. It is the work, written down. Our earlier piece on third-party vendor risk covers the security half; this is the commercial half of the same review.

The test grades your IT provider too, and it should. A managed services agreement is software procurement with a person attached. The same questions apply: is the price published, can you see the contract before you commit, what happens to your systems and documentation if you leave, and who owns what gets built. If a framework only works when pointed away from the person handing it to you, it is not a framework. It is marketing.

What Organizations Should Watch For

• An export that is a conversation. If getting your own data out requires a ticket, a fee, or a call with someone whose job is retention, the answer to the question you were asking has already arrived.
• Records held behind an active subscription. There is a real difference between paying for software and paying for access to your own history. Ask directly what happens to the data if you stop paying, and get the answer in writing, because the sales answer and the contract answer are not always the same.
• The upgrade that unlocks something that already worked. New capability behind a higher tier is ordinary product design. A capability moving up a tier after you built a workflow on it is a repricing of your switching costs, dressed as a release note.
• A published price with an unpublished renewal. Transparency about the first year and silence about the fifth is a specific choice. It is not disqualifying, but it tells you which of those two numbers the vendor would rather you compare.
• Terms you cannot read until you are nearly committed. If the agreement only appears after the proposal, the pricing, and the meetings, the document was sequenced to be signed rather than read.
• Anything sold to you as certified, guaranteed, or mandatory. Not a pricing tell, but it travels in the same company. Urgency and authority are the two things a strong product does not need to borrow.

The Honest Limits of This Test

A framework that only ever confirms what you already suspected is not a framework, so here is where this one stops.

Failing a check is not proof of bad faith. A small vendor may require an annual commitment because predictable revenue is what keeps it alive, and that vendor may still be the best thing available to you. Proprietary formats sometimes exist for genuine technical reasons rather than to trap you. Some products are expensive because they are expensive to build and support, and price level is not what this test measures.

A vendor can also pass on one axis and fail on another. Published pricing and a hostile export is a common combination, and so is the reverse. And a tool can score badly here and still be the correct choice, because it does something nothing else does, or because the cost of moving genuinely exceeds the cost of staying. The point of the test is not to produce a verdict. It is to make sure you are choosing the trade-off rather than discovering it in year three.

Recommended Actions

• Run the export test this week. Pick your single most important system, export everything, and open the file. Do not schedule it. Do it, because the friction you meet is the finding.
• Put every renewal date on one calendar. Add a reminder ninety days ahead of each. That window is the only point in the year when you have any leverage, and it passes silently for most practices.
• Ask what happens to your data if you stop paying, in writing. One sentence, by email, to every vendor holding anything you would miss. Keep the replies. They are vendor-oversight documentation whether or not you ever need them for anything else.
• Compare your invoice to the public price. If there is no public price, note that. If there is one and yours is different, find out what the difference is buying.
• Count licensed seats against people who actually log in. Departures often leave paid seats behind, which is a small bill and a standing access problem at the same time. Our piece on offboarding as a security event covers the other half of that.
• Find out who owns what you create. Templates, configurations, documentation, reports. If the agreement is silent, that silence is the answer, and it is worth resolving before it matters.
• Point all of it at your IT provider, including us. The contract, the price, the exit, the ownership. If you want help translating what you find, that is what managed IT is supposed to include, and the assessment is a reasonable place to start.