YouTube SEO

YouTube SEO services for a platform that is a recommendation engine first

Keywords get a video found. They do not get it distributed. Most views on YouTube come from suggested and browse rather than search, and those are earned by what happens after the click, not by what you put in the description.

Optimised for the two things the algorithm actually responds to: whether people click, and whether they stay.

The basics

What are YouTube SEO services?

YouTube SEO services make videos findable and, more importantly, distributable: researching what people search for on the platform, writing titles and descriptions that match it, designing thumbnails that earn the click, and shaping the video itself so viewers stay long enough for YouTube to start recommending it to people who were not searching at all.

Two audiences worth separating. Ranking inside YouTube is one job. Getting your videos to appear in Google results, where video carousels sit above ordinary listings for a lot of how-to and review queries, is a second and often easier one.

The core problem

You are optimising the smaller half

Most YouTube SEO advice concerns search: keywords in the title, keywords in the description, tags, a transcript. That work is real and it is worth doing, and it addresses the minority of how views are actually generated on the platform.

The majority arrives through suggested videos and the home feed, which are recommendation surfaces rather than search results. Nobody typed anything. YouTube decided your video was a good next thing for a particular viewer, and it decides that from behaviour: did people click when it was shown, and did they keep watching once they did.

The practical consequence: keyword work determines whether a video can be found at all, which matters most in its first days when it has no behavioural data. After that, retention and click-through decide how far it travels. A perfectly optimised description cannot rescue a video people stop watching at twenty seconds.

This is why channels sometimes find one video doing enormous numbers while technically identical uploads do nothing. The difference is rarely the metadata. It is usually that one of them held attention and the algorithm responded to that by showing it to progressively larger audiences.

The two levers

Click-through and retention work as a pair

These two metrics decide distribution together, and optimising either one alone tends to make things worse.

Click-through rate

Whether people click when your video is shown. Driven almost entirely by the thumbnail and title working as a single unit. A thumbnail that raises a question the title answers, or the reverse, outperforms two elements saying the same thing twice.

Audience retention

Whether they stay. The opening is decisive: if the first thirty seconds do not confirm the video will deliver what the thumbnail promised, the curve collapses and the video stops being recommended.

The failure mode of chasing click-through alone is well known and self-correcting in the worst way. A thumbnail that overpromises gets the click and loses the viewer immediately, which produces a high click-through rate attached to poor retention. The platform reads that combination accurately and stops showing the video, so the tactic works for about a day.

The retention curve is the most useful diagnostic you have. A cliff in the first thirty seconds means the opening is not delivering on the promise. A steady decline is normal. A rise partway through means something there is worth more than what preceded it, and that is usually a lesson about how the next video should open.

The other opportunity

Ranking your videos in Google, not just in YouTube

For a large set of queries, particularly how-to, setup, repair, review and comparison searches, Google shows a video carousel high on the page. Competing for a slot there is frequently easier than competing for the same query with a written page, because far fewer businesses produce video for it.

The work is different from ranking inside YouTube. It benefits from clear chapters so Google can surface a specific moment, an accurate transcript, a title that matches how the query is phrased in Google rather than on YouTube, and the video embedded on a relevant page of your own site with proper video markup.

Worth checking before producing anything: search your target queries in Google and see whether a video carousel appears. If it does, that is a route onto page one that your written competitors are probably ignoring. If it does not, the video is for YouTube and should be planned for YouTube.

Process

How our YouTube SEO services run

Find the demand

What your audience searches on YouTube, which queries show video carousels in Google, and which topics your channel has any chance of competing for given its current size.

Package before producing

Title and thumbnail decided before filming, because they define the promise the video has to keep. Deciding them afterwards is how videos end up with an opening that does not match the packaging.

Optimise the upload

Description, chapters, transcript, end screens and playlists, plus the metadata that helps Google index the video for its own results.

Read the curve, then iterate

Retention and click-through reviewed per video, with the specific moment viewers leave identified and fed into how the next one opens.

Deliverables

What our YouTube SEO services include

Platform keyword research

Researched on YouTube rather than borrowed from your web keyword list, because people phrase things differently when they expect a video answer.

Title and thumbnail packaging

Designed as one unit, with alternatives to test. This is the single highest-leverage thing on most channels and it is usually done last and quickly.

Opening structure

Guidance on the first thirty seconds, where retention is won or lost, based on where your existing videos actually lose people.

Descriptions and chapters

Written for viewers first and indexing second, with chapters that let both YouTube and Google surface the moment that answers a specific question.

Transcripts and captions

Accurate captions rather than automatic ones, which improves accessibility, watch time on muted autoplay, and what search engines can read.

Playlists and end screens

Structured so one view becomes two. Session length matters to the platform, and a viewer who continues watching your channel is worth more than one who leaves.

On-site video markup

Videos embedded on relevant pages with proper structured data, so they can appear in Google video results and support the page they sit on.

Retention analysis

Curve-by-curve review identifying exactly where viewers leave, turned into specific instructions for the next video rather than general encouragement.

Back catalogue work

Repackaging older videos that have decent retention and poor click-through. Frequently the fastest win available, because the hard part is already filmed.

Honest answer

When YouTube is not worth it for you

You cannot publish consistently

YouTube rewards sustained output. A handful of videos a year will not build a channel, and optimisation cannot compensate for a lack of material to optimise.

Nobody searches your topic in video

Some subjects are read rather than watched. Check whether video results appear for your queries at all before committing to production costs.

The production cost does not clear

Video is the most expensive content format per unit. If the same budget in written content would reach more buyers, that is the better trade and we will say so.

You want views rather than customers

Broad, entertaining topics get views. Narrow, boring, high-intent topics get customers. Optimising for the first and reporting the second is a common and expensive confusion.

One check worth doing on your existing descriptions and scripts: our readability checker will tell you whether your spoken script is written the way people talk, which is closely related to whether they keep listening.

Investment

How much do YouTube SEO services cost?

Optimisation is priced separately from production, because most channels have more videos than they have well-packaged ones and the cheapest gains are usually in what already exists.

Channel audit and back catalogue

$1,200 to $3,000 once. Retention analysis across existing videos, repackaged titles and thumbnails for the ones worth rescuing, and a keyword map for what to make next.

Ongoing optimisation

$1,200 to $3,000 a month. Research, packaging and upload optimisation for a regular publishing schedule, with retention reviewed per video and fed forward.

With production

$3,000 upwards a month. Where we are involved in scripting and structuring the videos themselves, which is where retention is actually decided.

If a budget only covers one thing, spend it on packaging and openings rather than on descriptions and tags. The metadata is cheap to fix later. A video that loses its audience in the first thirty seconds cannot be fixed later at all.

Common questions about YouTube SEO

Very little. They play a minor role in helping YouTube understand a video, particularly where a term is commonly misspelled, and they are nowhere near the lever people treat them as. Time spent on tags is almost always better spent on the thumbnail, the title and the first thirty seconds, which is where distribution is actually decided.

It depends entirely on where the video is being shown, which makes benchmarks less useful than they look. A video surfaced to a broad, cold audience will show a lower rate than one shown to subscribers, and both can be healthy. The number worth watching is your own, over time, on comparable videos, and always alongside retention rather than on its own.

As long as they hold attention and no longer. Longer videos can accumulate more watch time, which helps, but only if people stay. Padding a five-minute answer to twelve minutes produces a retention cliff that costs more than the extra runtime gains. Let the topic set the length and watch where the curve drops.

Yes, and accurate ones rather than the automatic captions. They improve accessibility, help viewers watching without sound, and give both YouTube and Google reliable text to work from. It is one of the few remaining metadata jobs that is clearly worth the time it takes.

Yes, and it is often the easier of the two. Google shows video carousels for a lot of how-to, setup, repair and review queries, and far fewer businesses produce video for those than produce articles. It needs chapters, an accurate transcript, a title phrased the way the query is phrased in Google, and the video embedded on a relevant page with proper markup.

Plan for consistent publishing over six to twelve months rather than a number. Early videos have no behavioural data, so they rely on search and on your existing audience, which is why channels feel stuck at the start. Once a video earns suggested placement the curve changes sharply, and that usually happens well before the channel is large.

Usually the fastest win available. Look for videos with decent retention and weak click-through, because those are ones people enjoy that too few are choosing to watch. A new thumbnail and title on that video costs an afternoon and the expensive part, the video itself, already exists and already works.

Find out where your viewers are actually leaving

We will look at the retention curves on your recent videos, identify the moment people drop, compare click-through against what the packaging promises, and tell you whether the problem is what you make or how you present it. Often several existing videos are worth rescuing. No obligation, no sales pitch.

Get my free channel review
  • Retention diagnosed
  • Packaging reviewed
  • Back catalogue triaged