YouTube is now the default search engine for how-to queries among UK users aged 18 to 44, and the data backs the shift: YouTube draws roughly 48.6 billion visits a month, making it the second most-visited site on the planet and the second-largest search engine after Google itself. Around 43% of UK consumers already use social platforms for everyday search, and up to 29.5% of Google AI Overviews now cite a YouTube video, making it the single most-cited domain in AI answers. The practical upshot for UK businesses is that one well-made tutorial can rank in three places at once: YouTube Search, Google's video carousel and AI Overviews. This guide shows you how to win all three. Expect production budgets from £400 to £2,500 per video, and a realistic 8 to 12 week ramp before a new channel earns consistent search impressions.
Last updated: June 2026
YouTube is replacing Google for how-to queries because people would rather watch someone do a task than read about it, and the numbers show that preference is now mainstream rather than fringe. Around 43% of UK consumers use social platforms for daily search, and 34% use video-led platforms such as TikTok daily. Among younger cohorts the tilt is sharper still: roughly 41% of Gen Z bypass traditional search engines and turn to a social or video platform first. eMarketer puts it bluntly, reporting that 46% of Gen Z prefer social platforms over search engines and that Google usage among Gen Z is down around 25% compared with Gen X.
For "how-to" intent specifically, video has a structural advantage that text cannot match. If you are searching "how to bleed a radiator", "how to set up a GoHighLevel pipeline" or "how to register a limited company in the UK", a video gives you the exact sequence, the visual cues, and the confidence that you are doing it correctly. Text forces you to translate words back into actions. Video removes that translation step.
Our honest view: this is not a fad, and treating YouTube as "social media we post to occasionally" is the mistake we see most UK businesses make. YouTube is a search engine with a watch-later layer bolted on. The query box at the top is doing the same job as Google's, and increasingly it is the box people reach for first when the answer is a procedure.
The behavioural shift breaks down into three drivers worth understanding before you spend a penny on production:
Here is the comparison that matters for a UK business deciding where to invest. The table below sets the consumer reality side by side.
| Signal | Google text search | YouTube search |
|---|---|---|
| Best for | Quick facts, comparisons, definitions | Procedures, demonstrations, walkthroughs |
| UK daily social/video search adoption | Still dominant overall | 43% use social search, 34% use video search |
| Gen Z first port of call | Declining (around 25% drop vs Gen X) | 41% turn here first |
| AI Overview citation share | Mixed domains | Up to 29.5% cite YouTube (top domain) |
| Content shelf life | Months to years | Often years; evergreen tutorials compound |
If your customers ever ask "how do I", a meaningful and growing share of them are typing that question into YouTube, not Google. That single fact reframes your content strategy.
YouTube shows up inside Google in three distinct ways, and this convergence is exactly why "winning both" is realistic rather than wishful: a strong video can earn a place in Google's video carousel, a standalone video result with key-moment timestamps, and a citation inside an AI Overview. Up to 29.5% of Google AI Overviews already cite a YouTube video, making YouTube the most-cited domain in AI answers overall. When Google's own AI generates a summary, it frequently points the user to a video to actually complete the task.
This is the strategic core of the article, so it is worth being precise about the mechanics. Google does not treat YouTube as a separate planet. Because Google owns YouTube, its crawlers read your video's title, description, transcript and chapter markers, then decide whether to surface the video against a written query. A query like "how to connect a domain to Cloudflare Pages" can return a written featured snippet, a video carousel and an AI Overview that cites a tutorial, all on the same page. One asset, several entry points.
The convergence also extends beyond Google. ChatGPT, Perplexity, Gemini and Copilot increasingly cite video sources, especially for procedural questions, because a transcript is just text they can read and a video is corroborating evidence they can point to. This is the under-covered angle most "YouTube SEO tips" lists ignore entirely. If you are optimising for answer engines (AEO) and generative engines (GEO), video transcripts are first-class citizens, not afterthoughts.
Here is how a single tutorial can surface across the modern search landscape:
The honest rule here: you do not chase six placements separately. You make one excellent video, optimise its metadata and transcript properly, and let the same signals satisfy every surface. The placements are downstream of the same fundamentals.
The ranking factors for YouTube and Google overlap on relevance and authority but diverge sharply on engagement: YouTube weights watch time, audience retention and click-through rate far more heavily than Google ever weights time-on-page. Understanding where the two systems agree and where they part is the difference between a video that ranks and one that disappears.
YouTube's algorithm is, at heart, a satisfaction-prediction machine. It is trying to guess which video will keep a given viewer watching and coming back. Google's algorithm, for written content, is trying to guess which page best answers a query and is most trustworthy. Where they meet is intent matching and authority. Where they split is what counts as a "good outcome": for YouTube, sustained attention; for Google, a satisfied click that does not bounce back to the SERP.
The side-by-side below is the comparison most generic guides skip. Keep it pinned when you brief any video.
| Ranking factor | Weight on YouTube | Weight on Google (for the video) |
|---|---|---|
| Keyword in title | High | High |
| Audience retention / watch time | Very high | Indirect (signals quality) |
| Click-through rate (thumbnail + title) | Very high | Moderate |
| Transcript / closed captions | Moderate | Very high (this is what Google reads) |
| Video chapters / timestamps | Moderate | High (powers "key moments") |
| Description depth (250+ words) | Moderate | High |
| VideoObject schema on an embedding page | Not applicable | High |
| Channel authority / topical consistency | High | Moderate |
| Engagement (likes, comments, shares) | Moderate to high | Low |
Notice the asymmetry. The transcript and chapters matter more to Google than to YouTube, because that text is precisely what Google's crawler ingests to understand the video. Retention and CTR matter more to YouTube, because they directly predict viewer satisfaction. To win both surfaces you must optimise for both columns at once, which is entirely doable because the levers do not conflict.
A stance worth stating plainly: chasing raw view counts is a vanity trap. A tutorial with 4,000 views and 65% average retention will outrank, and outlast, a flashy video with 40,000 views and 22% retention. Retention is the signal that compounds. Be sceptical of any agency that reports views as the headline metric and buries retention in a footnote.
The practical engagement targets we brief our clients towards in 2026 are:
You research keywords for video by prioritising long-tail, intent-rich phrases of three to five words over high-volume head terms, because video search rewards specificity and procedural intent. The phrase "how to add a custom domain in GoHighLevel" will convert and rank far better than the broad term "GoHighLevel", even though the broad term has more searches. On video platforms, intent beats volume almost every time.
The method we use is deliberately low-tech at the start, because YouTube hands you its own demand data for free. Type the start of a how-to query into YouTube's search box and read the autosuggest dropdown. Those suggestions are real queries people are typing, ranked by popularity. They are the most honest keyword tool that exists for video, and they cost nothing.
From there, a disciplined process looks like this:
The table below shows how to turn a broad seed into rankable targets. This is the work that separates a channel that grows from one that stalls.
| Broad seed term | Long-tail video target | Intent type |
|---|---|---|
| CRM | How to set up a sales pipeline in a custom CRM | Procedural / commercial |
| Chatbot | How to train an AI chatbot on your own FAQs | Procedural / informational |
| Automation | How to automate appointment reminders with workflows | Procedural / commercial |
| Odoo | How to configure Odoo inventory for a small UK shop | Procedural / commercial |
| Mobile app | How to test a mobile app before submitting to the App Store | Procedural / informational |
Our blunt advice for any new channel: ignore search volume for the first ten videos. Pick queries you can answer better than anyone currently ranking, build retention and authority, then graduate to higher-volume terms once the channel has earned trust. Volume without the right to rank is a slow road to nowhere. If you want this done at scale, our team can map an entire video keyword cluster as part of a broader business process automation or content programme.
You optimise a video to rank by getting six on-asset elements right: a keyword-front-loaded title, a click-worthy thumbnail, a 250-plus word description, a clean transcript, timestamped chapters, and an opening that answers the core question within 30 seconds. Get these right and you have satisfied the ranking factors for both YouTube and Google simultaneously.
Let us take each in turn, because the detail is where most videos fail.
Use five or more words and put the primary keyword near the front. "How to Set Up GoHighLevel Pipelines (Step by Step)" beats "GHL Pipeline Tips" on both relevance and click appeal. The title is the single biggest lever for click-through rate, and CTR is one of YouTube's strongest ranking signals.
This is the most undervalued asset in the entire chain. Custom thumbnails lift views by around 154% compared with auto-generated frames. A thumbnail needs three things: a clear focal point, large legible text of three to five words, and high contrast that reads well at phone size. Never ship an auto-generated thumbnail for a video you actually want to rank.
Write at least 250 words. The first two lines should restate the query and the promise, because that is what shows above the fold. The body should naturally include your keyword and its variants, a short summary of what the viewer will learn, and timestamped chapters. Google reads this text, so treat it like a mini blog post.
Treat a clean transcript as mandatory in 2026, not optional. The transcript is the primary text Google reads to understand a video and decide whether to surface it against a written query. Upload an accurate caption file rather than relying solely on auto-captions, which still misread technical terms and UK place names.
Add timestamped chapters in the description. Chapters power Google's "key moments" feature, which deep-links searchers to the exact second that answers their sub-question. They also improve retention by letting viewers navigate, which signals quality back to YouTube.
Answer the core question immediately. Do not open with a 45-second channel intro. Someone who searched "how to bleed a radiator" wants to see the bleed key in the first frames. Front-loading the answer protects retention, which is the metric that compounds.
Here is the optimisation checklist we run before publishing any client video.
| Element | Standard to hit | Why it matters |
|---|---|---|
| Title | 5+ words, keyword front-loaded | Drives CTR and relevance |
| Thumbnail | Custom, 3 to 5 word text, high contrast | Up to 154% more views |
| Description | 250+ words, keyword + chapters | What Google crawls |
| Transcript | Accurate uploaded captions | Mandatory for Google embedding |
| Chapters | Timestamped, query-aligned | Powers Google key moments |
| Opening | Core answer within 15 to 30s | Protects retention |
| Length | 8 to 15 minutes for tutorials | Room for depth without bloat |
On length: 8 to 15 minutes is the sweet spot for a how-to or deep-dive in 2026. Long enough to be genuinely useful and to accrue watch time, short enough to keep retention healthy. Pad a four-minute answer to twelve minutes and your retention graph will punish you. Match length to the genuine depth of the task.
You win both YouTube and Google with one asset by treating a single video as the spine of a content cluster: publish the video, turn its transcript into a written blog post, embed the video on that post, and add VideoObject schema so Google can read and rank it. This dual-funnel approach is the genuine "win both" framework that most articles promise and almost none deliver.
The principle is simple. You created the hardest asset, the video. The transcript is already 80% of a blog post. Reusing it costs a fraction of producing new content and multiplies your search surface area. Here is the full workflow we use.
The payoff is that one production effort now competes across multiple surfaces at once. The table below maps the asset to each placement it can win.
| Surface | Asset that wins it | Primary signal |
|---|---|---|
| YouTube Search | The video | Retention, CTR, relevance |
| Google video carousel | The video | Transcript, chapters, relevance |
| Google blue links | The blog post (from transcript) | On-page SEO, depth, links |
| Google AI Overview | Video + post | Clear atomic answers, schema |
| ChatGPT / Perplexity / Gemini | Transcript + post | Crawlable text, citations |
This is also where the AEO and GEO angle pays off, and it is the part competitors leave on the table. Answer engines cite procedural content heavily. A clean transcript and a well-structured post give ChatGPT, Perplexity and Gemini exactly what they need to quote you as the source for "how do I" questions. Video that lives only on YouTube, with no transcript-derived text, is invisible to most of these engines. Make the text exist and you become quotable.
Our honest take: the businesses that win the next three years of search are not the ones producing the most video, they are the ones producing the most reusable video. Build once, publish everywhere. If you want the technical scaffolding for this, from schema to embeds to internal linking, our AI automation agency and web application development teams build the pipelines that make repurposing near-automatic rather than manual drudgery.
You should choose video when the query is a procedure, a demonstration or a comparison that benefits from being seen, and choose a written post first when the query is a quick fact, a definition or a deeply technical reference people will scan rather than watch. Most businesses get this backwards, defaulting to text for everything and then wondering why their how-to content underperforms in 2026.
The decision is not video versus text as a religion. It is matching format to intent. Some queries are screaming for video; some are better served by a page you can Ctrl+F. The honest practitioner picks per query, not per dogma. Below is the decision matrix we apply.
| Query type | Lead format | Reason |
|---|---|---|
| "How to [do a physical or software task]" | Video first | Demonstration beats description; YouTube carousel appears |
| "How to [configure / set up a tool]" | Video first, post second | Screen-record then transcribe to a post |
| "What is [concept / definition]" | Post first | Quick scan; featured snippet potential |
| "[Tool A] vs [Tool B]" | Both | Video for the walkthrough, table-rich post for scanners |
| "[Technical reference / spec lookup]" | Post only | People scan and copy, they do not watch |
| "Best [service] in [location]" | Post first | Local intent, comparison, schema-led |
A few principles sit underneath the matrix. First, if Google already shows a video carousel for the query, that is Google telling you it wants video; follow the signal. Second, anything that involves watching a sequence of steps, where the order and the visual cue both matter, is a video candidate. Third, anything people will return to and copy-paste from, such as code, specifications or pricing, belongs in text where it is scannable.
There is also a budget reality. Video costs more to produce than text. If your resources are limited, prioritise video for the handful of high-intent procedural queries that drive your business, and serve the long tail with written posts. A small business does not need 50 videos; it needs the right 8 to 12 covering the procedures customers actually search before they buy.
Our view, stated plainly: for any business whose customers research "how to" before purchasing, video is no longer optional for those specific queries. The text-only competitor is ceding the carousel, the AI Overview citation and a growing share of younger searchers. You do not have to abandon writing; you have to stop pretending procedural intent is best served by paragraphs alone.
The Softomate implementation process turns the dual-platform strategy above into a repeatable production system, delivered in five stages over a typical 6 to 10 week initial sprint, with video content packages starting from £1,800 and ongoing monthly content programmes from £950 per month. We do not just give you tips; we build the keyword map, produce the videos, repurpose them into posts, add the schema and wire the internal links so each asset earns its full search footprint.
We are a London-based agency in Stanmore (HA7), and we run this as an engineering project, not a guessing game. Here is how the five stages break down.
We work to fixed quotes, not open-ended hourly billing, so you know the cost before we start. The timeline below sets realistic expectations for an initial engagement.
| Stage | Timeline | Deliverable |
|---|---|---|
| Discovery and keyword mapping | Week 1 | Approved video cluster + query map |
| Content architecture | Week 2 | Per-asset spec: video, post, schema, links |
| Production and optimisation | Weeks 3 to 6 | Optimised videos published |
| Repurposing and schema | Weeks 5 to 8 | Posts, embeds and VideoObject schema live |
| Measure and iterate | Weeks 8 to 10+ | Performance report + refinement cycle |
Indicative pricing for 2026: a single fully optimised video plus its repurposed post and schema starts at around £1,800; a 6-video initial sprint sits in the £8,500 to £12,000 range depending on production complexity; an ongoing monthly content programme that keeps the cluster growing starts at £950 per month. Where the strategy connects to automation, such as auto-publishing posts from transcripts or routing video leads into your CRM, that overlaps with our GoHighLevel automation services and custom CRM development work, and we will quote those elements separately and transparently. If video search is where your customers are heading, you can talk to us directly via our contact page.
Yes. YouTube attracts roughly 48.6 billion visits a month and is the second most-visited website globally after Google. By search query volume it is widely cited as the second-largest search engine, particularly strong for how-to, tutorial and product-demonstration intent where users prefer watching to reading.
Aim for 8 to 15 minutes for a tutorial or deep-dive in 2026. That gives enough room to genuinely answer the query and accrue watch time without padding. Match the length to the real depth of the task: never stretch a four-minute answer to twelve, because weak retention will hurt your ranking.
Effectively yes. The transcript is the main text Google reads to understand and rank a video against written queries, and it is what answer engines like ChatGPT and Perplexity can quote. Upload accurate captions rather than relying on auto-captions alone, which still misread technical terms and UK place names.
Aim for 7% or higher from impressions. CTR is one of YouTube's strongest ranking signals and is driven mainly by your thumbnail and title. Custom thumbnails alone can lift views by around 154% versus auto-generated frames, so treat the thumbnail as a first-class deliverable, not an afterthought.
Yes, and that is the whole point of the dual-platform approach. A single optimised video can appear in YouTube Search, Google's video carousel, standalone Google video results and AI Overview citations. Repurpose its transcript into a blog post with VideoObject schema and you also compete in Google's blue links.
Closely. Up to 29.5% of Google AI Overviews cite a YouTube video, making YouTube the single most-cited domain in AI answers. For procedural questions, Google's generated summary frequently points users to a video to actually complete the task, so ranking a tutorial can earn you an AI Overview citation too.
Match the format to the query. Procedures, demonstrations and software walkthroughs favour video first; definitions, technical references and scannable comparisons favour text. If Google already shows a video carousel for your query, that is a clear signal it wants video. Ideally you produce both from one production effort.
Usually 8 to 12, not 50. Focus on the high-intent procedural queries customers research before buying. A tight cluster of genuinely useful tutorials, each repurposed into a post, outperforms a large library of thin videos. Quality and topical consistency build channel authority faster than raw volume.
Because retention predicts viewer satisfaction, which is what YouTube's algorithm optimises for. A video with 4,000 views and 65% average retention will typically outrank and outlast one with 40,000 views and 22% retention. Views are a vanity metric; retention is the signal that compounds over time.
VideoObject is structured data you add to a page that embeds a video, giving Google the title, description, thumbnail, upload date and duration in machine-readable form. It makes the video eligible for video rich results and key moments. Yes, add it to any page where you embed a video you want Google to rank.
The shift is settled, not speculative: with 43% of UK consumers using social search daily, 41% of Gen Z reaching for video first, and up to 29.5% of Google AI Overviews citing YouTube, procedural how-to intent now lives substantially on video. The good news is that winning is not a choice between platforms. Make one optimised video, hit 7% click-through and 50%-plus retention, transcribe it into a 250-plus word post, embed it, add VideoObject schema and cross-link, and the same asset competes across YouTube Search, Google's video carousel, the blue links and AI Overview citations. Budget £1,800 for a single fully optimised asset, plan a tight cluster of 8 to 12 videos rather than fifty thin ones, and prioritise the procedures your customers research before they buy. The businesses that build reusable video, not just more video, will own the next three years of search. Start with the queries that matter most and compound from there.
If your customers search "how to" before they buy, it is time to claim both search surfaces. Explore how our London AI automation agency turns video into a multi-platform search engine for your business, or tell us your goals on our contact page.
Written by Deen Dayal Yadav, Founder of Softomate Solutions, a London-based AI automation and digital content agency in Stanmore (HA7). With over 12 years building software, automation systems and search-led content programmes for UK businesses, Deen has helped organisations move from text-only SEO to dual-platform video search strategies that rank on YouTube, Google and AI answer engines alike. Softomate Solutions is registered at Companies House, and you can read more about the team and our approach on our about page.
We protect the real names of all clients featured in examples and case studies. Every testimonial is from a real client.
More from the blog
Work with us
Book a free 30-minute discovery call with DD and get a personalised automation roadmap.
Softomate Solutions Limited
Deen Dayal Yadav
Online
We use essential cookies to keep the site running. With your permission, we also use analytics cookies to understand how visitors use our site so we can improve it. No data is sold. Privacy Policy
