How to build a pilot campaign.

I know nothing when I take on a new client who’s never used Google Ads for lead generation.

• I don’t know which keywords are going to convert.
• I don’t know what ad copy will resonate with the searchers.
• I don’t know which devices, demographics, times or locations will have the highest conversion rate or lowest cost per action.
• I don’t know how the competition will react. They might go to war to keep impression share or they might be asleep.
• I don’t know if search partners will generate any meaningful volume. Nor do I know if the CPA on search partners is going to be the same as search network or if it'll be cheap or exorbitant.
• I don’t know how well the client will respond to the leads. I don’t know if their sales process or offer is good enough to influence strangers to buy from them. I don’t know how long it’ll take from the time a potential customer fills in the enquiry form to the time they actually buy.
• I don't know how comfortable it'll be to work with the client.

I’m not saying I’m completely ignorant. I’ve done keyword research, looked at the competitor’s ads and websites and talked about the sales process and value of a lead. I'm building on many year's experience generating more than 2 million leads using Google Ads.

But, the truth is it’s all theory until the rubber hits the road. The only way to know for sure if Google Ads will work and if the relationship with the client will be comfortable is to test it.

A pilot campaign is the test

It's the smallest campaign that will answer these questions:

• How many sales leads can we get every month from Google Ads?
• How much will it cost to generate a sales lead?
• How well will these leads fit the business?
• How many leads will turn into sales?

A pilot campaign has to be comprehensive enough to stand a fair chance as a test, but simple enough to be built in a few hours.

Let me be clear here, when I say it must be simple I don't mean some junk thrown together without thought. It requires skill, care and diligence, the same as any other campaign. The difference - you're building less.

Choosing keywords for a pilot campaign

I'm looking for enough traffic to generate between 50 and 100 leads in about two months. If I can get a CTR of 10% and a conversion rate of 10% I need 100 impressions to get one lead. So I'm looking for 5 000 to 10 000 impressions in two months or 2 500 - 5 000 impressions a month. (More on the thinking behind this number here.)

10% CTR and 10% conversion rate is achievable with a pilot campaign in some markets. In other markets that's hard going. If I expect lower CTR or conversion rate I'll adjust the desired number of impressions upwards.

In the best case this traffic would come from searches for one service or product offering. It would have a high sale value and a short sales cycle. The client would be experienced at selling it. I want the pilot to be as easy.

This is not always possible. But I'd work very very hard not to run a pilot for a low margin offering, or one that most often took several months from enquiry to commitment.

I'll add extra offerings to the pilot if the keyword research shows that there isn't enough volume from one product.

Buying a high value product or service is a process that starts with research and ends with committing to a supplier. The words and phrases people use when searching give clues to how close they are to the buying end of the process.

For example, someone Googling for back pain is likely at the research end. They know they have a problem, but they haven’t honed in on a solution yet.

Someone searching for physiotherapist in newport is much closer to buying. Same with physiotherapist for back pain. physiotherapist for back pain in Newport is gold. These are high-intent or bottom of the funnel keywords.

I use high-intent keywords for the pilot campaign because people close to the buying end are more likely to complete an enquiry form.

Here's how I do it.

1. I start with a long list of possible keywords from our keyword research tool. 

2. I exclude one-word terms and most of the two-word terms. These short keywords are seldom worth advertising on.
3. I exclude any keywords that indicate that the searcher doesn't want to buy . I look for words like “how to”, “diy”, “ideas” etc.
4. I keep keywords that include locations eg “physiotherapist newport”. Keywords that include "near me" have also been good on some campaigns.
5. I keep keywords with buy signals like “price”, “buy”, “cost” etc. 


That leaves me with a bunch of keywords that aren’t obviously high-intent and can’t be excluded as obviously low-intent. I use a combination of experience and guesswork to choose enough to make up the search volume I’m looking for.

I’ll end up with somewhere between 70 and 300 keywords.

I start with the exact match keywords enabled. This gives me the most control over what searches trigger the ads. If I can’t get the expected number of impressions from exact match only I’ll enable broad match keywords.

(Pre 2019 I used to use phrase match. Google changed exact match to be much more broad so for now I've settled on broad or exact. I suspect this will change again but keeping up is part of our jobs :))

Negative keywords for pilot campaigns

I start with a list of about 200 standard negative keywords for lead generation campaigns. The list includes words like job, career, image, download etc.

I add negative keywords that I’ve spotted while doing keyword research.

Once the campaign is live I'll police the search terms report for negative keywords at least once a week.

Ad groups

Ad groups have only one purpose in a pilot campaign: give the the searcher a relevant advert.

I group keywords that are identical except for a few unimportant words. For example, the keywords used harley davidson softail classic and used harley davidson softail classic for sale differ only by the words "for sale". In my mind, the words "for sale" aren’t materially important in this search.

An ad reading something like ...

Used Harley-Davidsons for Sale - Softail Classics in Stock.
Contact us for great deals on used Harley-Davidson Softail Classics.

... would be a good match for either keyword, so I’m comfortable putting them in the same ad group.

In some cases this means that an ad group will have only one keyword. That's not a deliberate SKAG tactic.

Ad copy

I don’t A/B test ad copy in a pilot campaign, I write one ad per ad group.

• A campaign suffers no long-term damage if you delay split testing till you know that you can get leads at a reasonable price, and that the client can turn those leads into sales.

• I prefer to test significant changes rather instead of simple ones like swapping headline order or changing capitalization. (You should read this article on local maxima if you've only ever tested small changes.) The time spent producing a second and third set of good but completely different ads is be better spent elsewhere.

• You seldom get enough data in a short campaign to establish a conversion rate or CPA winner.


(More on my thinking about not testing here if you're interested: https://pete-bowen.com/google-ads-needs-deep-data).

Call to action / conversion method

I offer only an enquiry form on the landing page. If the client provides an emergency service eg locksmithing, plumbing, repairs etc I'll add click to call buttons.

I track conversions using Google Ads conversion tracking. I count one conversion per enquiry form with a 30 day conversion window. I don’t link Google Analytics during a pilot campaign unless the client is very familiar with it and has it producing meaningful data elsewhere.

Location

If my client serves locations where the level of competition is likely to be different - eg several large cities or multiple countries - I'll choose one for the pilot campaign.

Location is set to "People in your targeted locations". I don’t want ads shown to people outside the area my client can service.

Ad schedule

Ads schedule is set to show ads during time when the client can respond quickly - usually business hours if it’s B2B.

Other than that, I’ll adjust the other targeting settings from their defaults if I have a solid basis for doing so.

I’ve written about how I launch a pilot campaign in more detail here.

Related articles

Why a Google Ads account audit Isn't enough

Testing Google Ads for an RV Campground

Why negative keywords are becoming less effective in Google Ads

What’s the right bidding strategy for a brand new Google Ads account?

You can't run profitable Google Ads if your pipes are leaking

The campaign looks good but the ads aren't profitable

Topics you'll find on this site