Module 6 - Budgets

What do you need in place before you spend anything on YouTube ads?

Everything, before you get to spend a penny. That is the real key: making sure you have set up your conversion tracking, you have Google Tag Manager installed, you have optimised your landing pages, you have your remarketing loops built, and you have attached Google Analytics and your YouTube channel to Google Ads.

All of that is what I would call being ads ready. Until you have done all of those steps you cannot begin to spend any money, because if you do you are wasting your time and you are wasting your money. You are not going to learn and Google is not going to learn. I have described it as getting ready to set fire to a full bin of cash. Unless you are going to be bothered to set all the conversion events up, build the tracking in and make sure you are able to do your remarketing properly, do not bother.

Budgets are an absolute minefield of discussion in YouTube in terms of where you start. We get questions like how quickly can I scale, what is the minimum I can spend, how do I get success really quickly, how long does it take to learn. You need to fast forward this process, because otherwise the real danger is that you spend money really quickly and get nothing back. You get views, no clicks, and costs going through the roof. If you are out of control when you are setting up budgets at the start, you can spend fortunes. I have seen people spending thousands per week and just blowing budget without really knowing where they are going.

Should you start with one campaign or several?

It depends entirely on your daily budget, but the first thing to make sure you are not doing is putting all of your eggs in one basket. The temptation is to get going and set up one campaign.

If you only have a really small budget, let us say something under ten pounds or fifteen pounds per day, that is where you are going to have to start, with one campaign. That is fine. If that is what you have got to spend, stick with one campaign, go with what you think your best audience is going to be, go from there and learn, and take it steady.

The temptation, once you start getting some sort of results, is just to throw budget at it. If you do not follow this structure you get to the point where you set fire to money, and then you send an email to someone like me saying what have I done wrong. Normally, nine times out of ten, putting all the eggs in one basket is where most people go wrong, because with a limited budget you really need to be able to spread it and see what happens.

Why should different audiences go into different campaigns?

Because if you put three or four audiences into one campaign with one budget, Google will spend all the budget on one audience. Effectively you have done the all eggs in one basket thing again, because Google just spends the money on the one audience it could serve, which is probably the broadest and probably the one you wanted to test the least.

Let us say you have a budget of something up to maybe fifty pounds a day. It is far better to split that budget into three or four campaigns and test different audiences. When people start, they tend to add multiple audiences to a single campaign. What we would much rather you do is split those audiences into different campaigns.

If you want to combine them, maybe put two into one campaign, so if you had four campaigns you could test up to eight audiences, two in each. The chances are Google will still spend most of the budget on one of the audiences within each campaign, so that is the reason we split them out into multiple campaigns rather than having one campaign with multiple audiences inside it.

The whole point of running multiple campaigns is so that you can split test different audiences. You may be testing different creative too, but quite a lot of the time what we do is try maybe two or three different videos, the same ones in each campaign, with different audiences. If you have got the resource to split test different adverts across different audiences, you are going to learn even more.

Bear in mind some of them may well not serve. If they do not serve, you have not wasted your budget and you can test again elsewhere. If you throw all the budget into one campaign, the chances are one audience will swallow it all.

What structure works for a budget of around fifty pounds a day?

Split it into four campaigns of twelve pounds fifty each. That is a good little structure that I like to follow to start with.

Google probably will not spend it all. You will probably find that maybe one or two of your campaigns do not even serve anything to begin with, especially if you are going to start on maximum cost per view. If you want to spend the budget, which you probably do in the early stages because you want to learn, then setting maximise conversions as the goal means Google will spend your budget. It will attempt to spend everything you give it.

So when you are testing in the early stages, if you have four campaigns spending twelve pounds fifty a day, you have spent fifty pounds a day or thereabouts. What we are trying to do is see what a conversion is going to cost us and what our cost per view for certain placements is. When you run on a max conversion strategy, it is going to show you within the campaign different costs for different placements. You will be able to see where your conversions are coming from, albeit they are micro conversions, things like time on site as an initial.

Why start with time on site as a conversion event?

Because Google then will not send you really poor quality traffic that just bounces off your site. We want to find traffic that comes on and actually sticks.

You might, for example, use a conversion event of time on site of thirty or sixty seconds. At least then we know people are coming on and staying and looking at our content, so they have got a chance to move to the next conversion up. That may well be viewing a particular video or product, or adding something to the cart, or making a phone call or an enquiry. That is where we are looking to get people to next.

In other modules of the course you will have seen that we teach you to start with the micro conversions and phase them out as we get more of the conversions we actually want, the ones that are more valuable. Be careful not to optimise your campaign for time on site forever. We teach Google what we actually want and what we are prepared to pay for it.

When do you move from maximise conversions to target cost of acquisition?

Once you can get to fifty or more conversions in a single campaign inside thirty days, again even if those are micro conversions like time on site. Within the first thirty days we want to look at moving over to a cost per click strategy, rather than sticking on max conversions, so that we take the reins from Google and force Google to give us better conversions at a better cost, or more conversions at a better cost.

Once you understand that Google is going to give you a conversion for a certain figure, let us say you are getting conversions for six pounds, Google should be able to deliver and improve on that if you started on a max conversion or max cost per view model. What we are now telling Google is: you have done your job, I now want you to do better, I want you to improve, and I am prepared to go with target cost of acquisition, but the most I am prepared to pay is x.

How high should you set your target cost of acquisition?

Two to three times what you expect to get it for to begin with. That might sound massively complicated, but if you are getting conversions at about six pounds you need to give Google some headroom to start with so it can learn.

You have just gone from a strategy where you are saying please maximise my conversions, Google, and you see conversions at six to ten pounds. When you move to a target cost of acquisition model, start at roughly two and a half times that. Two and a half times six is fifteen, so at fifteen you would tell Google I am not prepared to pay more than fifteen pounds for a conversion. Set your target CPA up at that and see if Google achieves it or betters it.

This is where the logic comes in. People would think, right, I have seen conversions at six, I will go in and set it at five. What probably happens in that scenario is that all your campaigns stop serving altogether and you sit there scratching your head going, what went wrong, Google, you told us you are already doing six, surely you can do better. In reality you have to give Google a bit of room. You have to ask, can I achieve this, and Google says right, okay, I am prepared to go with twelve pounds or fifteen pounds.

Then slowly over time we move that target cost of acquisition down. To do that we remove non-performing placements from the campaign, to teach Google this is where you are going wrong.

What is the objective in the first thirty to sixty days?

To learn, not to profit. The objective over that first thirty to sixty days is to learn the conversion data. Let Google start to learn, and you learn your audiences and learn what is working.

We are not necessarily looking for profit within this period. It would be nice to get straight away into positive figures, obviously, but you are investing in the campaign at this point.

How do you structure a budget of around 250 pounds a day?

Split it into ten campaigns of twenty-five pounds a day. Half of those would initially be on maximise conversions and the other half set on a max cost per view, so we are looking to try and get people onto the site for a certain cost.

Again, we want to move over to a target cost of acquisition once you have got fifty conversions in a particular campaign inside a thirty-day period, and then you can let Google learn, because it will serve more people to you at the same cost. Start low, and the objective again is to learn the audiences and begin to learn where that conversion data is.

How do you work out what your cost per view should be?

Take the cost per click your competitors are paying in traditional PPC and divide it by fifty. If you are in a particular business it is a good idea to look at the cost per click that your competitors are paying and bidding to get people onto the site in traditional PPC. You can find that out when you search keywords in Keyword Planner. You can see the bids and you can see what people are bidding to get to the top of normal PPC.

If somebody is paying a pound to get a click onto a site from traditional PPC, that is a text based advert, and you divide it by fifty, it takes you to two pence. That is where I would want to target my cost per view. What I am basically saying is, can I get two per cent of the people to click through to my site having shown them a video? If you have shown somebody a video and they have skipped it at up to thirty seconds it is free anyway, so we are not bothered. Of the people who have gone past thirty seconds, can we get two per cent of those to click through onto our site, or at worst one per cent of them?

That would mean we would be doing better than a traditional PPC campaign. We would be getting people to land on the site at a more affordable cost, and they have watched a video, so they know more about our product, they know more about us, and we have built up more trust.

Conversely, if somebody is paying ten pounds a click in a particular industry, you probably do not need to go as high as dividing that by fifty to get to a cost per view. You might do, but I would probably still start with a three, four or five pence cost per view and see if your campaign serves. If it does not, you have got headroom in it.

What does it take to spend more than 250 pounds a day?

Enough data in your existing campaigns to know that it is going to work. If you want to come into YouTube ads and spend more than 250 pounds a day, you had better have that data. You cannot just come in at 250 a day. Or you can if you want, and if you have got the budget you could come in at a thousand or more a day, but you will not be scaling properly.

At this point you should have built up to having all of your campaigns on target cost of acquisition, and then you are able to lower those conversion values as you go forward. In the example I gave before we had a target cost of acquisition of fifteen pounds for an enquiry, and now we are talking about proper conversions, not micro conversions.

This is a massively complicated maths equation at this point, but what we are effectively saying is: Google, I am prepared to pay x for a sale, or I am prepared to pay y for an enquiry. These are valuable to me, and this is how we teach Google to do it. We then try to lower those values and see where campaigns start to lose traction or audiences start to shrink.

How often should you relaunch and split test campaigns?

We split test these campaigns and republish on a bi-weekly or monthly basis. Bear in mind that every time you relaunch a campaign it is going to start learning again, but if we have got enough conversion data from a similar audience, Google will be able to replicate what it has done before quickly.

This is what you need to do to spot the patterns that Google produces within your particular industry, and this is something that cannot be taught. I would love to teach you how your industry reacts. Unfortunately I do not have the ability or the knowledge or all the data to say, right, here are five thousand different industries and products and a video on every single one of them to show how Google reacts to different audiences and products and landing pages. That is over to you.

You have got to learn that, and you have got to learn how often you need to republish, because it depends on how many competitors you have got, how competitive the bids are, which placements you have got, which audiences, everything. The trick is to see where campaigns stop performing, and you will see patterns. I do teach that later in the course, so pay attention to that.

Those are three basic strategies. Once you get to stage three and you are spending 250 a day or more, you need to have it all on target cost of acquisition, because it gives you that much more control. That is not to say you cannot go testing more strategies with smaller pieces of budget and newer audiences, because you will have loads and loads of conversion data by this point. That is how you scale it.

Hopefully that has been useful, and I will see you in the next little segment.