Forecasting your future paid search performance can be a difficult task, but is a fundamental skill that all digital and performance marketing managers need when determining how much of your media budget to distribute to your channels and projecting revenue figures for senior management.
This guide will help you acquire the data that you need to answer questions like the following:
- What would happen if we increased or decreased budgets by £X
- What would happen to leads volume if we could spend £7 to acquire a lead as opposed to £5
- What would happen if we increased bids by x%
- What would happen if we had 100% impression share for these keywords
- Can you create a 12 month forecast for our paid search activity
This guide will mainly focus on forecasting for accounts that have historical data, as opposed to forecasting where you are using Google’s Keyword Planner to get estimates for keywords that you are planning to put into your account.
Forecasting what would happen if you increased or decreased budgets.
One of the easiest ways to forecast performance is by using the Google Performance planner tool that was rolled out on May 14th 2019. It’s a brilliant tool for forecasting different scenarios and what the impact of them on your account.
The Google performance planner tool can be found within the tools and planning section of your Google Ads account.
You first set a forecast period, for example the next 12 months. A metric that you’re aiming towards such as generating conversions or conversion value, and then finally you can set a target conversion value / CPA / Spend depending on what you’re trying to forecast.
Once you have selected your date range and metrics you should then get a graph that looks like the following
The gray dot shows the current CPA and spend levels. You can then move the blue dot along the line to determine how many conversions you would get for your budget.
Or alternatively if your senior management team asked what would happen if we set a $10 CPA as opposed to a $7 CPA how many more conversions could we purchase?
You can also download the plan and see campaign by campaign breakdown of spend, conversion and CPA.
One of the limitations of the tool is that you can’t forecast month by month very easily which your upper management may be interested in seeing. Here you would have to create a new forecast for each month using the tool.
As an alternative to Google’s performance planner you can use regression models to see how CPA would vary based on volume.
This has already been covered in this guide to using regression analysis to forecast Google Ads performance so I won’t cover it here.
Forecasting what would happen if you increase or decrease bids
Keyword level bid simulators are a great way to determine what would happen if you were to increase or decrease bids by a certain percentage.
To view them you should view your keywords within Google Ads and add all of the bid simulator columns. It will allow you to see what would happen to performance if you were to either decrease bids by 50% or increase them by 50% or 300%.
You won’t always want to increase bids by 50% you might want to know what would happen if you were for example to increase bids by 30% instead.
There are also limitations as Google only shows you what happens to clicks and cost. You may want to determine what would happen to your conversion volume, your cost per acquisition or ROAS for example.
To do that you will need to use your excel skills and follow these steps.
Step 1. Once you’ve added all of the bid simulator columns to your keyword view, download the report into Excel. You can use Google’s filters to only select the keywords that you want to get forecasts for.
Step 2. Total up the columns so that you can see the total cost and clicks at the current point in time. Then calculate the total costs and clicks for -50% bids, +50% bids ands +300% bids by adding the totals of those columns to the current cost and clicks.
Step 3. Plot the points on a graph. This will allow you then to determine the increased cost and increased number of clicks that you would expect to get.
Step 4. Cost and clicks might not necessarily be the metrics that you would like to forecast. For example if you wanted to forecast Conversion volume and conversion cost.
You would take the number of clicks you expect at each data point (-50%, +50% and +300%) and times it by the conversion rate that you currently have as conversion rate does not usually vary by position that much.
You can then determine the number of conversions. To determine the CPA then take the new adjusted cost at each data point (-50%, +50% and +300%) and divide that by the number of conversions. You can then plot them on a graph as I have done below.
You can then apply the same methodology to get any of the metrics that you would like.
Forecasting what would happen if you had 100% impression share
For brand protection campaigns, having a 100% impression share is key to success. Especially in competitive markets where competitors are likely to bid on your brand.
To determine what would happen if you have 100% impression share you need to add impression share columns to your campaigns report and filter the report by the campaigns that you want to make the forecast for.
You should then look at the percentage of impression share that is currently being lost. For example if you have an impression share of 85% you are losing 15%.
Other metrics like conversions and revenue can be projected fairly easily by multiplying the number of current conversions or revenue that you have by the percentage difference between your current impression share and 100% impression share.
This is quite simply a rough way to make forecasts and will only really work for branded terms. As it relies on two main premises.
Firstly that you currently show position 1 for your branded terms in paid search when your ads show so you won’t see an increase in CTR when you give the campaigns more budget.
It secondly holds on the premise that as you currently show position 1 and by increasing impression share you will appear more for position 1 and therefore your conversion rate will not vary.
Putting it all together to forecast a 12 month paid search media plan
As part of your overall media plan you may want to create a 12 month paid search forecast that will be included in part of your wider media buying plan. And potentially create further forecasts split by Search vs Shopping vs Display for example. Or Brand vs non-brand vs competitors.
When you’re finished you should have a document that looks something like this.
Step 1. Create a new Google Sheet with the 12 months down the side (as this is the format that Google exports reports in) and then add column headers for the metrics that you want to see as you can see in the example above.
Then name this section last year and then create an identical section below labeled this year.
Step 2. Select the last 12 months data that you have within Google Ads. Then head to Reports > Predefined Reports > Time > Month and you should see a month by month breakdown of your performance. Download this as a report in Excel format.
Step 3. Copy that data and drop that into the spreadsheet in last years data section.
Step 4. This step will vary depending on what you are looking to achieve.
If you’re looking to maintain your current position you can ask Google for your industry trends report that will show you how much CPCs and search volume has increased over the last 12 months naturally and then apply these numbers to your forecast so you can see roughly how many more impressions / clicks you will get naturally over the next 12 months and at what cost.
If you’re looking to increase / decrease your budget, target a specific CPA or ROAS then you should follow the steps to use Google Performance Planner to forecast the performance for each month.
If you’re looking to maximise your impressions share on brand and forecast out performance then follow that section of the guide.
When working on a 12 month plan, it’s generally a good idea to keep the budget splits per month from the year before the same to take into account seasonality.
When it comes to forecasting what will happen with increased or decreased budgets Google’s new performance planner is a great tool. It’s also effective for modelling what will happen to CPA and conversion volume at different budgets as an alternative to creating your own regression model.
When it comes to modeling what happens when you increase or decrease bids. Google’s bid simulator is an effective tool for providing forecasts of how performance would vary at different bid increases.
Modelling based on impression share is also relatively simple provided that you’re working with branded terms by simply looking at the percentage difference between the current impression share and 100% impression share and forecasting based on increasing the current figures by this difference.
Finally when asked to create 12 month paid media forecasts, you should be able to put together the three forecasting types that we have covered in the article to create projections for each month based on your targets and overall media budget.