Module 5 - Where to start

Where should you start with YouTube ads?

Start with placement campaigns if your budget is small, because they keep you in videos and channels where your content is highly relevant. It is often the case that clients come to us saying, how are we going to set campaigns up to start with? That very much depends on what your objectives are, short and long term. Obviously the objective is as many conversions as we can get on a reliable basis, and growth over a period.

If you are starting off with a reasonably small budget, and this will not apply to all of you, anything probably up to fifty pounds a day, you will probably want to start with placement campaigns. To learn how to do that, have a look at the Vidnetics option within this course. It will show you how to harvest placements. You can also harvest placements based on your topics or keywords from within the setup, and there is a module within the course to do that.

So the best place to start if you have got a budget up to fifty pounds a day is to run campaigns based on placements only, and see how it goes. See if you are getting impressions, see if you get conversions.

Why is a broad keyword a problem on YouTube?

Because Google knows far more about the person than the keyword tells you, and it will place your ad wherever that person happens to be watching. If I am searching today for golf clubs, Google knows what type of person I am, what I do, what my hobbies and interests are, what other types of websites I visit, and which videos and subjects I watch on YouTube.

The issue is that if you use a broad match, or just the keyword golf club, that can match an awful lot of people. Even if you targeted guys in their forties, golf club is very broad. So you could find me watching a video about DIY, and you could find another guy similar to me who had also searched golf clubs watching a video about cooking.

The problem with that is the placements Google is selecting are not necessarily matching for the rest of the people watching those channels. You have just chosen, for example, Fred's DIY channel, and I am watching a video on that. If you serve the ad into that channel again, it is not necessarily people who are like me. They will have matched your criteria, in that they will have searched golf club, but does that placement then work for that particular individual? Are there even enough people watching that channel who match your criteria?

So cleaning your placement list is going to be essential. Again there is a module within the course on how to set exclusions, learn about your placements, and learn where your conversions are coming from. This is a massive learning process to get YouTube ads to run properly.

Why can you not go far wrong with a placement campaign?

Because the person is already searching for the topic and is already watching a video that matches your keywords. If you are starting out with a small test budget, start with placements. They are looking for a video with probably your keywords in the title, so they are watching a DIY video. If you have got a DIY product, slot it into a DIY video and you have got a decent chance.

You would then overlay other demographics such as age, or income if you are in the US, and other parts of the targeting. You might overlay some keyword targeting into that as well.

How do you test multiple audiences without wasting budget?

Set lots of little campaigns up with small budgets rather than one large one. Once you get to wanting to scale a bit further, you are going to want to test multiple audiences.

If you use maximise conversions as the target for the campaign, what Google will do is maximise conversions, which basically in Google language stands for please spend all my budget. So where we are trying to test multiple audiences, if you are going to use a max conversions strategy, try setting lots of little campaigns up with small budgets.

If I was going to go past fifty pounds a day spend and I now wanted to run several different tests with different audiences, the way to do that is choose max conversions when you get going, because your account to start with does not have any conversions in it. We need to find out where they are and let Google learn. The easiest and safest way to do that is to test multiple audiences with small budgets. Yes, it will spend your budget within each campaign, but hopefully you can quickly learn without spending too much budget which of your campaigns are going to win.

When do exact match keyword campaigns come in?

They come next, once you have learned from your audience tests, and they are for reliability rather than speed. These will take longer to learn and they will be slower to grow, but they should be more reliable in the long term in showing you which keywords are going to work for you.

This is still the early stages, before we get to the further stages of scaling and running campaigns that are highly targeted. If we are running keyword campaigns with six to twenty keywords in an ad group, you can then find out which of those keywords work well and which do not, and that is going to help you build future audiences.

When should you switch to max cost per view?

Switch once you have learned where your conversions are coming from and which keywords or audiences work. Stepping on from keyword campaigns, you are going to want to build some custom audiences, and set these types of campaigns up starting with a max cost per view rather than maximise conversions.

So set, for example, three, four, five or even ten pence per view as your maximum. You will have to test to see where that sits. Start off low, as I have explained in other parts of the course. Start off small. Does it deliver any impressions? If not, copy the campaign, put the cost per view up, let it run again, and see if it serves and finds some impressions for you, and obviously then finds conversions.

It is all about learning, testing and finding what is working. The key part from this video is that you need to find those successful placements, you need to find which keywords are working, and you need to find where your ads work and why. You will only understand that through putting ads out, putting ad sets out, and seeing where time on site, form fills, phone calls and purchases all come from, and why, and within which audiences.

When are you ready for smart bidding and a target CPA?

You are ready once you have a conversion set that is working and you know how many conversions you get and what they cost. At that point Google will ask you whether you want to spend more for an extra number of conversions, and you can move to Google smart bidding and a target cost of acquisition.

When you start out with a brand new YouTube ads account, Google will suggest a target cost of acquisition. I am saying to you now, you are better off using the strategy above: working with placements and your own audiences, testing your own assumptions, learning if they are right, learning if Google's off-the-shelf audiences are right, moving into some exact match keyword campaigns, and then getting a number of conversions into your account.

What is a micro conversion?

A micro conversion is a small signal of interest that has no sale value on its own but marks a better visitor. It might be just getting somebody to a landing page where they stay on the site for over a minute. That does not actually have any true value in terms of a sale or an enquiry, but the value is that the person who does that is more important to you than the person who lands on the site and then disappears.

If that is your first conversion event, and let us say we get a thousand people onto the website and 300 of them stay for more than a minute, Google can then optimise for more people who stay. Then once we get more people who stay, that leads on to more people who make an enquiry or purchase a product. So we are asking Google to give us more people of the same type and find us more conversions at our target cost per acquisition.

Should you set a target CPA on a micro conversion?

I would not necessarily, although I do see that happen with quite a lot of people. I saw somebody today who came to me with landing page view as their conversion event. What they had done to cheat was set the target cost of acquisition at a really low figure.

Typically within their search ads, their normal PPC, they were paying one to two pounds per click. On YouTube they were achieving lower than that by setting a target cost of acquisition for a landing page view of 50 pence, so half the cost. People were then having to watch 30 seconds of their video in the main before they clicked through. It was quite clever, but it was sort of gaming the system.

What are the basic suggestions, then?

Start with placements if you have got a small budget and you are just testing. There are so many ways you can set this all up, but that is where to begin. You can obviously run multiple placement campaigns based on different keywords. That can work really well and you can scale up that way. It does depend on whether you are able to find enough videos and channels that match your placement criteria.

Bear in mind when you are setting the placement campaigns up that you can choose other parts of the display network such as apps and websites, and not just YouTube and YouTube search. Other parts of the course will cover that off. Those are my basic suggestions for where to start, so I will see you in the next little module.