Disclosure: this page contains affiliate links. If you buy through one I earn a commission, at no extra cost to you. It has not changed the numbers below.
Every YouTube tool sells keyword research as the growth lever.
The screenshots always show a green score and a big volume number.
Both of those numbers are estimates, and both are widely misread.
YouTube does not publish search volume to anyone.
Every figure you have ever seen in a YouTube keyword tool is modelled. 🔍
That does not make the tools useless. It makes them useful for something narrower than advertised.
Here is what they genuinely know, what they guess, and how to use the output.
🎬 See Tube Magic’s Keyword Tool →
🧾 What a Keyword Tool Actually Knows
Separate the measured from the modelled before you trust a screen.
| Data point | Measured or guessed | How reliable |
|---|---|---|
| Autocomplete suggestions | Measured | Real, scraped from YouTube |
| Videos currently ranking | Measured | Real, visible to anyone |
| View counts of those videos | Measured | Real and public |
| Upload dates and channel sizes | Measured | Real and public |
| Monthly search volume | Modelled | An estimate, often rough |
| Competition score | Modelled | A formula, not a fact |
| Overall opportunity score | Modelled | A blend of the two above |
The first four rows are why these tools are worth money.
The last three are why two tools give different answers for the same phrase.
📉 Search Is Not Where Your Views Come From
This is the finding that changes how you use every keyword tool.
📉 Roughly where a typical channel gets its views
Why this matters so much
Keyword tools optimise for search. Search is usually the smaller half.
Chasing volume alone means polishing your third biggest traffic source.
The bigger lever is whether the topic makes people click at all.
Check your own numbers first
Open YouTube Studio, then Analytics, then Traffic Sources.
Read the split for your last twenty videos, not the site-wide average.
Tutorial channels skew heavily towards search. Entertainment channels barely touch it.
What to do with that split
If search is under a fifth of your views, treat keywords as topic input.
If search is over a third, keyword work deserves real effort.
Notre YouTube SEO guide covers the second case in detail.
🔢 How to Read a Volume Number
A practical translation table, because the raw figure misleads.
| Reported volume | What it usually means | Worth targeting? |
|---|---|---|
| Under 100 | Real but tiny demand | Only if it converts well |
| 100 to 1,000 | The sweet spot for small channels | Yes, usually |
| 1,000 to 10,000 | Established channels dominate | Sometimes, with a better angle |
| Over 10,000 | Crowded and defended | Rarely, early on |
| Zero or no data | Too new or too narrow to model | Check autocomplete instead |
The zero row catches people out constantly.
A missing estimate is not proof that nobody searches for it.
Why estimates run low on new topics
Models need history, and a two-month-old topic has almost none.
Emerging phrases show near-zero volume while demand is rising fast.
That is precisely when the topic is easiest to win.
The autocomplete cross-check
Type the phrase into the YouTube search bar and watch the suggestions.
Those suggestions come from queries real people typed.
If a phrase autocompletes, somebody is searching it.
🔎 Try Unlimited Keyword Lookups →
⚖️ Competition Scores Are Relative
Here is why two tools disagree about the same keyword.
⚖️ Two ways of scoring the same phrase
The question that actually matters
Not how hard the keyword is in general.
Whether a channel your size has ranked for it recently.
That is answerable by eye, in about ninety seconds.
How to answer it manually
Search the phrase on YouTube and look at the first ten results.
Note the subscriber count of each channel and the upload date.
If three results come from channels under 5,000 subscribers, the door is open.
The signal that says walk away
Ten results, all from large channels, all more than two years old.
Old videos with huge view counts are extremely hard to displace.
Pick a narrower phrase instead of fighting that page.
🎯 The Three Keyword Types Worth Targeting
Most phrases fall into one of these, and the difference is strategic.
| Type | Example shape | Pourquoi ça marche |
|---|---|---|
| Problem keywords | How to fix a thing | Urgent, specific, high intent |
| Comparison keywords | This against that | Viewer is close to deciding |
| Beginner keywords | How to start something | Endless, renewing audience |
| Vague topic keywords | A single broad noun | Avoid, no clear intent |
| Brand keywords | A product name | Only if you review products |
Problem keywords are the most reliable for a new channel.
Somebody with a problem right now will watch a plain video from a small channel.
Why intent beats volume
A thousand idle searches are worth less than a hundred urgent ones.
Urgency drives clicks, watch time and subscriptions together.
Volume alone predicts none of those three.
The comparison advantage
Comparison videos rank quickly because few people bother making them well.
They also attract viewers who are ready to act on advice.
That is why they earn more per view than general content.
🧪 Where the Three Tools Differ
The keyword features specifically, not the whole product.
| Tool | Keyword approach | Coût |
|---|---|---|
| Tube Magic | Unlimited lookups, generation focused | $47 monthly |
| vidIQ | Scores and daily ideas, credit metered | Free plan, then $16.58 |
| TubeBuddy | Scores plus bulk tag tools | Free plan, then a few dollars |
On keyword research alone, the free plans cover a lot of ground.
Tube Magic earns its price on generation, not on lookups.
What unlimited really buys
Freedom to research forty topics without watching a credit counter.
That matters when you plan a quarter of content in one sitting.
It matters very little if you publish twice a month (Futurepedia, 2026).
Where the free tiers stop
Credit limits bite when you research in bulk or generate scripts.
Until then they are usable products rather than crippled demos.
Notre vidIQ comparison lists the exact limits.
🚀 Research Without Credit Limits →
⏱️ A Twenty Minute Research Process
This is the whole method, and it fits on one screen.
| Minutes | Étape |
|---|---|
| 0 to 3 | Type your seed topic into YouTube autocomplete |
| 3 to 8 | Collect twenty phrases the bar suggests |
| 8 to 12 | Run them through your tool for rough volume |
| 12 to 16 | Search the best five and check who ranks |
| 16 to 19 | Pick the one where small channels appear |
| 19 to 20 | Write the title before you write anything else |
Step four is the one nobody does, and it decides everything.
Looking at the actual results page beats any score.
Why the title comes first
A title you cannot write compellingly is a topic you should skip.
Deciding this before filming saves you a wasted afternoon.
Notre title generator guide explains that test.
Why twenty minutes is the ceiling
Research has sharply diminishing returns after the first pass.
Hours spent in a keyword tool are hours not spent filming.
Publishing more is the stronger growth lever at every channel size.
🚫 Five Mistakes That Waste the Effort
Each of these is common and each is avoidable.
| Mistake | What happens instead |
|---|---|
| Chasing the highest volume | You lose to bigger channels |
| Trusting one tool blindly | You inherit its model bias |
| Ignoring the results page | You miss the obvious answer |
| Stuffing tags with keywords | Tags barely matter now |
| Optimising an unclickable title | Ranking without any clicks |
The tag row surprises people who learned this years ago.
Tags carry far less weight than titles and thumbnails today.
Why tags faded
YouTube reads your actual spoken words and on-screen text now.
It no longer needs you to describe the video in a list.
Fill them in briefly, then stop thinking about them (OutlierKit, 2026).
What replaced them
The title, the thumbnail and the first line of the description.
Those three decide both ranking and clicking.
Spend your optimisation time there instead.
💷 Is a Paid Keyword Tool Worth It
An honest answer that depends on how often you publish.
| Votre situation | What to use |
|---|---|
| Under two videos monthly | Free plans and autocomplete |
| Two to five videos monthly | A cheap paid tier if you feel limited |
| Six or more videos monthly | Unlimited becomes worth paying for |
| Planning a whole quarter at once | Unlimited, clearly |
| Managing a large old library | Bulk tools matter more than lookups |
Six videos a month is roughly where the maths turns (Product Insight AI, 2026).
Below that, the free tiers genuinely do the job.
The break-even in plain terms
At $47 monthly and six videos, you pay about $7.83 per video.
If the tool saves an hour on each, that is a good trade.
Notre pricing breakdown works that through fully.
La recommandation honnête
Start free, publish more, and upgrade when limits interrupt your work.
Being interrupted is a real signal. Marketing is not.
That order costs nothing and rarely goes wrong.
📊 The Four Numbers That Predict a Video
Keyword volume is only one input, and it is the weakest of the four.
📊 How much each factor moves the result
Why click-through rate leads
A video nobody clicks cannot rank, however well optimised it is.
YouTube shows your thumbnail a fixed number of times to start.
What happens in that window decides everything afterwards.
This is why title and thumbnail work beats keyword work.
They are the only two things a browsing viewer ever sees.
Why retention comes second
Clicks get you shown. Retention keeps you shown.
A high click rate with poor retention stalls within a day.
The first thirty seconds carry most of that weight (Sanish Tech, 2026).
Where volume really sits
Volume tells you a topic exists. Nothing more than that.
It cannot tell you whether your version will be clicked.
Treating it as a growth predictor is the core mistake here.
🧭 Turning Keywords Into a Content Plan
Research is only valuable if it produces a queue of videos.
| Scène | What you produce | Time |
|---|---|---|
| Seed list | Twenty autocomplete phrases | 10 min |
| Filter | Eight with visible small channels | 15 min |
| Title test | Six you can title well | 15 min |
| Cluster | Two groups of three related videos | 10 min |
| Schedule | Six weeks of publishing | 5 minutes |
Under an hour gives you a month and a half of content.
That is the real return on keyword research.
Why clustering matters
Three related videos feed each other in the suggested column.
A viewer who finishes one is offered the next immediately.
Scattered topics never build that internal pull (Futurepedia, 2026).
How to build a cluster
Take one strong phrase and find two natural follow-up questions.
Publish them close together, then link them in end screens.
The middle video usually outperforms the other two.
Why six weeks is the right horizon
Long enough to stop scrambling for topics every week.
Short enough that your niche has not moved underneath you.
Replan when the queue drops to two remaining videos.
🔬 Testing Whether Your Research Worked
Most creators never check whether the effort paid off.
| Metric to check | Where | What good looks like |
|---|---|---|
| Search impressions | Traffic Sources | Rising over 28 days |
| Which queries found you | YouTube search terms | Terms you targeted appearing |
| Click-through from search | Reach tab | Higher than browse rate |
| Position of your video | Manual search | First page within a month |
| Views after 30 days | Overview | Still accumulating slowly |
The last row separates search content from feed content.
Search videos keep earning views long after publication.
The long tail effect
A feed video spikes and then dies within a fortnight.
A search video trickles for years without further work.
That difference compounds across a library of thirty videos.
Reading the search terms report
YouTube tells you the exact phrases that surfaced your video.
They are frequently not the phrases you targeted.
Those surprises are your best ideas for the next batch.
When to give a video more time
Search rankings settle slowly compared with feed performance.
Judge a search video at sixty days, not at seven.
Deleting it early destroys the asset just as it starts working.
🧩 What AI Adds to This Process
A fair account of where generation genuinely helps.
| Tâche | AI value | Remarques |
|---|---|---|
| Expanding a seed into phrases | Haut | Fast and genuinely varied |
| Grouping phrases into clusters | Haut | Tedious work done instantly |
| Drafting titles per keyword | Moyen | Needs heavy editing |
| Judging competition | Faible | Look at the results yourself |
| Deciding what you should make | Aucun | That is your judgement |
The top two rows are worth real money on a busy channel.
The bottom two are where people over-delegate and regret it.
The expansion advantage
Turning one topic into forty phrases is dull manual work.
A generator does it in seconds and finds angles you missed.
That alone justifies unlimited access for high-volume channels (Product Insight AI, 2026).
The judgement limit
No tool knows your audience, your voice or your back catalogue.
It will happily suggest topics you are wrong to make.
Filter every list against what you can film convincingly.
A practical division of labour
Let the tool produce breadth. Let yourself decide depth.
That split gets you the speed without the generic output.
It is also how the creators who use these tools well work.
💡 Verdict on AI Keyword Research
Plainly, and with the commercial caveat stated openly.
| Claim | Verdict |
|---|---|
| Volume numbers are accurate | False, they are estimates |
| Competition scores are objective | False, they are formulas |
| Keyword research helps | True, as topic selection |
| It is the main growth lever | False, titles and retention are |
| You need a paid tool | Only above six videos monthly |
I earn commission if you buy Tube Magic through this page.
I still think most readers should start on a free plan.
Who should pay for unlimited
Créateurs publiant chaque semaine ou plus, planifiant le contenu par lots.
A ce volume, les limites de crédit deviennent un véritable obstacle.
Les recherches et scripts illimités rapportent ensuite confortablement 47 $ (Orbit River, 2026).
Qui ne devrait pas
Toute personne publiant une à deux fois par mois, sans exception.
Les niveaux gratuits de vidIQ et TubeBuddy couvrent cela confortablement.
Payer maintenant achète une marge que vous ne toucherez pas (Busy Mom Side Hustle, 2026).
La seule habitude qui vaut plus que n’importe quel outil
Recherchez votre phrase sur YouTube et étudiez la première page.
Quatre-vingt-dix secondes de recherche battent un après-midi de tableaux de bord.
Chaque bonne décision de mot clé que j'ai prise a commencé là.
🗂️ Où s'inscrit la recherche de mots clés en une semaine
Un dernier élément de contexte, car le timing change la valeur.
| Day | Work | Rôle du mot clé |
|---|---|---|
| Journée de planification | Choisir les trois prochains sujets | Centrale |
| Journée d'écriture | Rédaction et structuration | Aucun du tout |
| Journée de tournage | Enregistrement | Aucun |
| Journée d'édition | Découpe et polissage | Aucun |
| Journée de publication | Titre, vignette, description | Touche légère |
Le travail sur les mots clés s'effectue sur un jour sur cinq au maximum.
Les créateurs qui dépensent plus que cela publient moins et grandissent plus lentement.
L'habitude de la journée de planification
Regroupez chaque recherche en une seule session tous les quinze jours.
Laisser l’outil ouvert toute la semaine invite à une remise en question constante.
Une file d'attente fermée de sujets décidés protège votre temps de tournage.
Le chèque du jour de publication
Confirmez que la phrase apparaît naturellement dans le titre et la ligne d’ouverture.
Alors arrête. Un bricolage supplémentaire change rarement quelque chose de mesurable.
Envoyez la vidéo et passez l’heure économisée sur la suivante.
❓ Questions fréquemment posées
Does YouTube publish real search volume?
No. Every number in every tool is a model.
Ils sont utiles pour classer les expressions les unes par rapport aux autres, et non comme une vérité absolue.
Why do two tools show different volumes?
Because they use different models and different data sources.
Traitez l'écart comme une plage de confiance plutôt que comme une erreur.
Is a zero volume keyword worthless?
Not at all. It often means the topic is too new to model.
S'il se complète automatiquement sur YouTube, de vraies personnes le tapent.
Do tags still affect ranking?
Very little. YouTube reads your audio and on-screen text directly.
Ajoutez-en quelques-uns pertinents et continuez.
Should I put the keyword in the title exactly?
Include it naturally, but never at the cost of a clickable title.
Un titre maladroit et exact est classé puis ignoré.
How many keywords should one video target?
One primary phrase, plus whatever related terms appear naturally.
Essayer d’en couvrir cinq dilue les cinq.
Does Tube Magic include keyword research?
Yes, with unlimited lookups on its single $47 plan (Orbit River, 2026).
Il n'y a pas de niveau gratuit, seulement une garantie de remboursement de 30 jours.
Is vidIQ enough on the free plan?
For most channels publishing a few times a month, yes.
Le niveau gratuit comprend des scores, des idées et des crédits IA mensuels.
How often should I redo keyword research?
Once per content batch, not once per video.
Le faire plus souvent produit à nouveau la même liste.
Does search volume matter more in some niches?
Yes, heavily. Tutorials, software and repair topics live on search.
Commentaires, vlogs et divertissements en direct sur les flux suggérés.
What single check improves results the most?
Looking at the real results page for your phrase.
Rien dans un tableau de bord ne remplace le fait de voir qui se classe actuellement.
Do keywords matter for Shorts?
Far less. Shorts are distributed almost entirely through the feed.
Écrivez pendant les deux premières secondes plutôt que pour une requête.
Can I rank without any tool at all?
Yes. Autocomplete plus the results page covers the essentials.
Les outils payants permettent principalement de gagner du temps une fois que vous publiez fréquemment.
📚 Références
IA d'analyse de produit. (2026). TubeMagic : qu'est-ce que c'est ? Avis honnête et tarifsConsulté le 22 août 2026 sur https://productinsightai.com/what-is-tubemagic-ai-review-pricing/
Futurepedia. (2026). Avis sur TubeMagic AI : cas d’utilisation, prix et alternativesConsulté le 22 août 2026 sur https://www.futurepedia.io/tool/tubemagic
Rivière Orbit. (2026). Analyse de Tube Magic 2026 : Outils d’IA pour la croissance de YouTubeConsulté le 22 août 2026 sur https://orbitriver.com/tool/tube-magic-ai-review-2026/
OutlierKit. (2026). vidIQ vs TubeBuddy : répartition des fonctionnalités et des prixConsulté le 22 août 2026 sur https://outlierkit.com/resources/vidiq-vs-tubebuddy/
Activité secondaire d'une maman occupée. (2026). Avis sur Honest Tube MagicConsulté le 22 août 2026 sur https://busymomsidehustle.com/tube-magic-review/
Sanish Tech. (2026). Avis sur TubeMagic : fonctionnalités, prix, avantages et inconvénientsConsulté le 22 août 2026 sur https://sanishtech.com/reviews/tubemagic-review/
Lectures complémentaires sur ce site
Notre Revue complète de Tube Magic couvre l'ensemble du produit. Voir aussi le growth tools guide et notre small channel guide. Voir aussi notre idea generator guide.
À propos de cette page
Cette page gagne une commission d'affiliation et recommande toujours des outils gratuits pour les chaînes publiant rarement. Les prix et les fonctionnalités ont été vérifiés le 22 août 2026 et peuvent changer, alors vérifiez sur les pages officielles avant d'acheter.
- AI Video Idea Generators: Turning 100 Ideas Into One Good One - septembre 2, 2026
- YouTube Automation Tools in 2026: What Can Actually Be Automated - septembre 1, 2026
- Subscription Creep: How Small Fees Become Big Bills - août 31, 2026