Collapsing anchors is an NLP technique that neutralises an unwanted feeling by firing its anchor at the same time as a stronger, resourceful anchor, so the two states blend and the negative charge fades. Step by step it runs in eight moves: define both states, test that the resource is the stronger one, build each anchor on a clearly separate spot, break state between them, fire both together, release the negative anchor first, hold the resource a few seconds longer, then test the original trigger. The order is not decorative. Almost every failed collapse I see in a training room comes from swapping two of these steps or skipping the test.
| Question | The short version |
|---|---|
| What is collapsing anchors? | Firing an unwanted state and a stronger resource state at the same time until the unwanted one loses its charge |
| How long does it take? | Ten to twenty minutes, most of it spent building the two anchors properly |
| How many anchors? | Two, on clearly separate spots, never on the same knuckle or the same shoulder |
| What is the key rule? | The resource must be stronger than the problem state, and it is released last |
| How do you test it? | Fire the old trigger, not by asking how the person feels |
| When is it the wrong tool? | Trauma, diagnosed conditions and anything that belongs with a clinician |
What does collapsing anchors actually do?
It gives one cue two competing responses and lets the stronger one win. An anchor is a stimulus linked to a state, the way a particular song drops you back into a summer. Fire two at once and the nervous system cannot run both states cleanly, so what comes out is a blend in which the resource dominates. The old cue is still there afterwards. What changed is what it produces.
The mechanism has a name outside NLP: counterconditioning, which is well studied in learning research. That literature is also where the honest caveat sits. Mark Bouton at the University of Vermont has spent decades showing that a conditioned response weakened in one setting can return in another, an effect he calls renewal. A collapse that holds in a quiet training room can wobble in the actual meeting room. That is an argument for testing in context and repeating, not against the technique. The glossary entry on collapsing anchors in NLP places the pattern in the wider model, and the entry on anchoring in NLP covers the groundwork this technique assumes.
Collapsing anchors step by step
Read this as a working script rather than a summary. Each step has a reason attached, because the reason is what tells you whether you can shorten it.
- Name both states precisely. One unwanted state tied to a real situation (“the tightness when my inbox shows his name”), and one resource state chosen because it is genuinely available to this person. Vague states produce vague anchors.
- Check the resource is the stronger of the two. Have the person rate both from one to ten, then get the resource above the problem before going further. A resource at four against a problem at eight does not blend, it gets swamped, and the session ends worse than it started.
- Build the resource anchor first. Elicit the state fully, with the body and the breathing, not as a memory described from outside. Anchor at the peak on one identifiable spot, for example the second knuckle of the left hand. Break state, then fire it alone to confirm it works.
- Build the problem anchor second, and lightly. You need enough of the state to work with, not a full re-run of a bad afternoon. Use a spot that cannot be confused with the first, for example the same knuckle on the right hand. Break state and test it once.
- Break state properly between every step. An unrelated question, standing up, a change of subject. Without this both anchors hold a mixture, and a mixed anchor cannot collapse anything.
- Fire both anchors at the same time. Hold them through the part that looks strange: uneven breathing, asymmetry in the face, a colour shift, sometimes a small laugh. That is the two states competing, and it usually lasts ten to thirty seconds.
- Release the problem anchor first, keep the resource five to ten seconds longer. This one detail decides the result. Let go of the resource first and the last thing the system holds is the problem state.
- Break state, then test with the trigger. Fire the old anchor alone and watch, then have the person think about the original situation. If the charge is gone, close with a rehearsal of the next real occasion. If it is not, the resource was too weak, and you return to step two.
The closing rehearsal is not optional, given what the renewal research predicts. Take the next concrete occasion, find the cue inside it, and run the new response there. The seven future pacing examples show what that level of detail looks like in practice.
What should you see while it happens?
You should see the body do something you did not ask for. That is the reason a practitioner watches instead of listens. During the simultaneous fire, the usual signs are breathing that goes briefly out of rhythm, one side of the face doing something different from the other, a colour change around the neck, and then a settling that arrives on its own.
- Before: a stable baseline you have watched for a minute, so you know what normal looks like for this person.
- During: asymmetry and irregularity. Do not interrupt it and do not talk over it.
- After: shoulders lower, jaw releases, breathing lengthens. Often a puzzled expression, because the old feeling did not show up on cue.
- Not a good sign: nothing at all, which usually means one anchor never held.
Reading these signals is a trainable skill and it is the bottleneck for most new practitioners. If your calibration is shaky, the anchoring exercises are worth running before you attempt a collapse on anyone, because a technique you cannot observe is a technique you cannot correct.
What makes collapsing anchors fail?
Six failures cover nearly everything. Five of them are procedural and cost nothing to fix once you can name them.
- The resource is too weak. The most common single cause. Build it up or choose a different one, and never proceed on hope.
- The two anchors are too close together. Adjacent spots on the same hand bleed into each other, and you end up firing one blurred anchor twice.
- No break state. Both anchors then hold a blend of both states from the beginning, which makes the collapse meaningless.
- The resource anchor is released first. A small slip of the hand that reverses the outcome of the whole session.
- The test is a question. “How do you feel now?” gets you politeness. Firing the old trigger gets you information.
- No rehearsal afterwards. The collapse holds in the room and gets no chance to hold anywhere else.
One limit is not procedural and deserves to be said plainly. If the unwanted state is a trauma response, a diagnosed anxiety disorder, an eating disorder or a phobia that shapes daily life, a pair of anchors is the wrong instrument. NLP work of this kind complements professional care and does not replace it. Refer, and say so out loud to the person in front of you. A trainer who does not teach you that boundary has left something important out of the course.
Can you collapse anchors on yourself?
You can, and it works less reliably, for a structural reason rather than a mystical one. Collapsing anchors depends on watching a body change while its owner is busy inside a state. Doing both jobs at once means the observer half keeps pulling you out of the state half, so the peak you anchor sits lower than you think.
If you do run it alone, use anchors you cannot confuse, hold the simultaneous fire longer than feels necessary, and write down in advance which hand you release first. Then test it against a real situation within a few days rather than against your own impression five minutes later.
In our trainings, the correction I make most often during this exercise has nothing to do with anchoring at all. People rush step two. The problem state is vivid and interesting, the resource state feels obvious, so they build the resource in thirty seconds and spend four minutes on the problem. The collapse then fails, and the participant concludes that the technique is unreliable rather than that the ratio was wrong. In groups of no more than ten I can see whose resource anchor is thin before they fire anything, which is the only reason this gets caught in time. Carolin Mallmann, Licensed Trainer of The Society of NLP, Lisbon
That is also why anchoring is taught live rather than from a handout. Over seven consecutive days of supervised practice in Lisbon, participants run collapses on each other repeatedly, watched and corrected, until the sequence and the observation happen at the same time without effort.
Frequently asked questions
How long does collapsing anchors take?
Between ten and twenty minutes for one pair of states, with most of that time spent building and testing the two anchors rather than collapsing them. The simultaneous fire itself lasts ten to thirty seconds. If a practitioner offers to do it in three minutes, they are skipping the elicitation, which is where the result actually comes from.
Where should you place the two anchors?
On two spots that cannot be mistaken for each other, such as one knuckle on each hand, or one knee and one shoulder. Distinctness matters more than location, because anchors placed close together blend before you intend them to. In a professional setting, ask permission before touching anyone and use a neutral spot.
What do you do if the old feeling comes back?
Treat it as expected rather than as failure. Learning research on renewal shows that a weakened response can return in a new context, so repeat the collapse and rehearse the new response in the specific situation where it is needed. If it returns at full strength every time, the resource state was too weak and needs rebuilding.
What is the difference between collapsing anchors and the fast phobia cure?
Collapsing anchors works by blending two states through simultaneous anchors, and keeps the person associated in the resource. The fast phobia cure works by distance, having the person view a distressing memory from outside itself. They are used for different intensities: a stuck feeling on the one hand, a phobic response on the other, and the second belongs with a trained practitioner.
Can you use collapsing anchors with more than one resource state?
Yes, and it is often the fix when a single resource is not strong enough. Stack two or three resourceful states on the same spot, one at a time, testing as you go, then collapse the stacked anchor against the problem anchor. This keeps the ratio in your favour without pushing the person into a state they cannot reach.
Keep exploring
- Anchoring exercises you can practise alone, the groundwork this technique assumes, if your anchors do not hold reliably yet.
- Seven future pacing examples, our most recent piece, for the rehearsal step that decides whether a collapse survives outside the room.
Carolin Mallmann is a Licensed Trainer of The Society of NLP, trained by Richard Bandler. She has trained more than 300 people and runs English-language NLP trainings in Lisbon, in groups of no more than ten.
