You searched “network effects” or “lock-in”, probably because someone told you: get big enough that users cannot leave, and you win. You want to know whether to race for it. This page does not teach you how to lock people in. It gives you one judgment first: people kept only because leaving is costly are a wall you built, not a place you earned. Ask three questions before you decide whether this ground fits you.
“They can’t leave” sounds like a moat. It looks more like entangling ground
In business today, network effects usually means a product gets more useful to each person as more people use it. That is a modern idea; Sun Tzu never said it. What you read here is this site setting it next to the Terrain chapter only to borrow one judgment.
The chapter says 可以往，难以返，曰挂, which this site’s English gives as “Ground that can be entered but is difficult to withdraw from is called entangling terrain.” If you lead users into a place that is easy to enter and hard to leave, that is entangling ground for them. The text goes on: there, if the other side is prepared and you attack without success, “it will be difficult to retreat, putting you at a disadvantage.” If you keep people only by making return hard, the day they find another place, you are holding ground that serves no one.
Check one: would people stay if leaving were free?
Run a thought test. Suppose that from tomorrow your users could take their data, contacts, and habits with them at no cost. Would they stay? If you can name the reason, such as the product truly works or the people they need are truly here, this ground may fit you.
If you cannot, and the only answer is “moving is a hassle”, admit it plainly: it is the wall keeping them, not you. This site states the chapter’s judgment directly: “The same force changes outcome when the ground changes. If the ground does not fit, do not fight.” When the wall drops, the outcome changes sides.
Check two: are you building a reason to stay, or a cost to leave?
List what you plan to ship this quarter and ask of each item: does this make people want to stay, or make it harder to go? The first is value; the second is cost. Not every cost is wrong; some moves are hard by nature. But you should know it is a cost you carry, not a prize you won.
The chapter says 夫地形者，兵之助也, given here as “Terrain is an aid to warfare.” Ground is an aid, not the whole. If you make switching costs your main force, you have turned the aid into the core. One check on the chapter page fits exactly: “Does this ground help you preserve, or force a hard fight?”
Check three: when growth only comes from higher walls, when do you stop?
Write yourself a stop rule: if the next stretch of growth comes mainly from making leaving harder rather than staying better, you stop and change posture. This site’s chapter page puts the next step plainly: if the posture does not match, change posture rather than force it. Do not treat winner takes all as a fight you must take by default.
The chapter also says a true commander “seeks only to preserve the state and the army”. Bring that home to your own case: your team, your users, the people willing to come back next round. Protecting them is harder to copy than fencing them in, and it is one less needless fight.
Questions people ask after searching “network effects”
Did Sun Tzu talk about network effects?
No. Network effects is a modern business term. Sun Tzu never discussed it, and this site does not present it as his idea. This page borrows only one judgment from the Terrain chapter: the same force changes outcome when the ground changes. If you read that Sun Tzu already explained network effects, that is a stretch. Keep the source text, the translation, and your modern reading apart.
Is locking users in always wrong?
Not always. Some things are hard to move by nature, such as long-built data or shared work habits, and that is not your fault. The question is what you treat it as. As a cost you carry, you will try to make leaving decent. As a prize you hold, you will keep raising the wall. Ask yourself: if users could take everything free tomorrow, would you still sleep well?
Rivals are all racing for scale. If I don’t, will I fall behind?
First ask whether this ground fits you. Scale is central to some businesses and only someone else’s question for others. If you race just because rivals race, they are moving you. Write down why your users would choose you a second time. If you can, put your effort there. If you cannot, scale you win will not hold.
How do I tell growth from value apart from growth from switching costs?
Look at the people who stayed recently and ask why. If the answers are “it works” or “the people I need are here”, that is value. If they are “switching is a pain” or “I can’t export my data”, that is cost. You can also count your own roadmap: how many items improve the experience, and how many raise the barrier.
Does winner takes all mean I have to fight to the end?
Not as this site reads it. Winner takes all describes a market; it is not an order. On distant ground, the chapter says that where both sides are evenly matched, “forcing engagement will be disadvantageous.” If you and a rival are evenly matched, a fight to the end usually hurts both. You can ask first whether this fight is necessary and what would make the next cycle better.
What should I do after reading this?
Write three lines: why people would stay if leaving were free; which items this quarter build reasons and which build costs; and your stop rule if growth only comes from higher walls. If you cannot write the first line, stop there. If you can write all three, open the Terrain chapter and set your answers against the four checks on the chapter page.
The next step is not another plan for locking people in. Ask whether people would stay if leaving were free, separate the work that builds reasons from the work that builds costs, and write a stop rule. That judgment is the one this chapter is named for on this site: Ground does not fit. If your team wants this check first, Contact Us. We talk teaching, training, and content, not locking users in. We reply within two business days.