Contact us

How to create a data-driven YouTube strategy7 min read

7 min read

We’re living in unprecedented times (the more I hear that phrase, the more I want to scream). But it also means more brands are looking to innovate and keep connecting with their customers. 

More brands are pivoting to YouTube as a way to connect with their audience during the coronavirus lockdowns. But that doesn’t mean it should be a knee-jerk reaction. You should still be thoughtful and measured when creating content for YouTube. 

Making data-driven decisions is one of the best ways to ensure that you have structure to your YouTube account and that you’re hitting the agreed KPIs. Let’s not forget that creating video content can be extremely expensive and time-consuming, so you should be aiming to get the most bang for your buck that you can.

Structure and order

The first thing to do is to think about how you will structure your categories and themes. Where do you want to appear? How will you organise who is creating what? 

Organising your YouTube content into distinct pillars is incredibly useful to assign KPIs, budget and resource. This is especially important for larger organisations, where different stakeholders (and budget pots!) are often needed across videos. 

We structure our YouTube strategies around three different “types” of content, based on the data that goes into informing the strategy:

1. Brand queries

This is the first space you want to enter as a brand, and it makes the most sense right? These are spaces where your brand should be appearing, and information that you, as the brand, should be giving your opinion or the relevant information to the audience.

This is slightly obvious, but brand searches are any search terms that contain your brand or any product names. 

KPIs for this type are:

  • Subscribers – people are already interested in the brand if they’re searching a branded query, so they should be easy to convert into subscribers
  • Referral traffic – similar to the above, it shouldn’t take much to drive users from a branded query to the main site, as long as there is a relevant and matching page on the site

2. Non-brand queries

Non-brand queries are videos that are aimed at expanding your audience and establishing you as an authority in other areas.

Non-brand queries can be anything that doesn’t contain your brand or product names. Bear in mind that non-brand queries can still contain a brand name, it just doesn’t have to be yours.

A good example of this is CNBC covering Levi’s. CNBC is capitalising on searches surrounding Levi’s, that Levi’s themselves have not covered.

 

KPIs for this type are: 

  • Subscribers – one of the main reasons to brand out into non-brand queries is to expose your brand to a new audience. Therefore, subscribers are a good measure of whether or not this is working 
  • Engagement – Subscribers tell you if people are interested in the brand, but engagement is also good to see how the audience feels about your brand weighing in on certain topics. If you start to see heavy amounts of Dislikes surrounding a particular topic or theme, consider pivoting away from this space. Do you have the authority to speak here? Perhaps not.

3. Campaign messaging

The last “type” of video is campaign messaging, which really means any old videos that are associated with a particular campaign. These don’t have to be TV, they can be anything that you have run in any space (Facebook, Instagram, Influencer, User-generated etc).

Pro Tip: Absolutely organise these into playlists so that people can see everything associated with one campaign in one place. Alternatively, group playlists by type (i.e. TV ads) to help your audience see how your brand has changed over the years

KPIs for this type are a lot looser than the previous two, as this “type” is mainly just a repository of old campaign videos. However, it’s still important to measure! So what we would suggest is measuring the following:

  • Views – this will tell you how much longevity a campaign has. If people are still watching a video months (or years) after a campaign has finished, then there must be something to it, right? In which case you might want to work with that particular talent again (either actor or individual). Or maybe the direction of the video was great? Did you incorporate comedy? If so, it might be worth doing this again. These are all important reflections on old campaign videos that are only unlocked if you host the video on YouTube post-campaign
  • Referral trafficReferral traffic is good to measure too. If people are still finding the ad that engaging that they’re heading back to the site, it’s definitely something to keep in mind planning out future campaigns of that nature

Creating a holistic approach

YouTube is essentially a search engine, so it makes sense to approach it like one. And like all good SEO strategies, you need to start with keyword research.

There are two tools you can use to do this:

VIDIQ

VidIQ is a super useful tool for both optimising videos and for research of new videos. It has a list of useful features, but one of the best features that will help you plan is the Volume and Competition metrics.

By entering a search term, you will be able to find out the Volume and Competition of any keywords and make an informed decision about where to start when creating content.

For each of the strategic pillars, a hierarchy of needs should be developed based on the core info that people want to know. By basing a roadmap around dominating areas we already know are popular, momentum can be gained, new subscribers can be added and the strategy can evolve.

By reviewing the results pages of various brand and non-brand queries, it’s possible to build a picture of what exists, how popular content is, who’s dominating and where the areas of opportunity are. 

VidIQ has a free trial/plan and a useful Chrome extension that you can use (I’m not on commission, I swear). 

Solomon

The second tool, and my preferred method, is to use Croud’s own proprietary technology, Solomon. Solomon is a dynamic content ideation and planning tool, powered by machine learning. Fancy, right? 

Solomon allows us to run keyword analysis for YouTube at scale. We are able to extract video titles, URLs and engagement metrics from YouTube, categorise these videos into themes and topics and then overlay other metrics; such as volume and difficulty. 

Plotting these metrics in a bubble graph is an easy way for us to visually see where the opportunities are for a client to begin to start to create content. 

Plotting data on engagement, competition and search volume together provides a clear view on which areas have the most potential. You can see a macro level the themes and topics that provide short-term strategic gains. 

The data that goes into each category is also captured so that we drill down into each of the topics, in order to inspire ideation and therefore know what specific titles or keywords are being used in videos. 

Solomon provides us with an easy way to do this quickly and efficiently at scale, analysing hundreds of thousands of videos in the space of a few hours. 

Once you have your keywords, you’re ready to go. There are some basic production rules to abide by, such as repeating the keywords in the audio of your videos, the same as you would with copywriting. But as with copywriting, quality is of the utmost importance when producing your videos. All the keywords in the world cannot help a piece of content rank that ultimately is of poor quality. 

Looking to dip your toe into the world of YouTube (or throw yourself in at the deep end) and need some support? Want some advice? Or maybe just tips? Let one of our lovely content experts know and we’ll be more than happy to help.