Chances are, by now you've heard about the benefits of long-tail keywords. Chances are, you're already using long-tail keywords in your Google AdWords campaigns.

But exactly how long-tail are your campaigns? Too little? Too much? Are your long-tail keywords actually adding any real value to your campaigns?

An effective long-tail Google AdWords strategy is not just a simple case of permutating lots of different themes together, to create thousands of keyword combinations. If you do that - you'll probably end up with millions of keywords in your campaigns, of which only a tiny percentage will ever get shown. Your account will become bloated, your campaigns will become difficult to manage, and all that time you spent in Excel will prove fruitless.

An effective Google AdWords strategy is instead about using real data to drive your campaign expansion in a strategic, intelligent, and methodical fashion. It's about starting small, using negative keywords to funnel desired traffic to desired campaigns and ad groups, and expanding keywords based on the themes that are most popular and beneficial to your business.

Tremendous challenge

Back in 2012 I wrote an article for SEP about the practicalities of creating large-scale long-tail Google AdWords campaigns. I realised that a huge challenge existed for location-based service businesses such as cleaners, landscapers, home builders, and paint stores. Here is that challenge:

"There are an infinite number of possible ways people can describe your products and services"

It might not sound like a big deal, but it really is. Think about it for a second:

Someone looking for carpet cleaning might search Google for "cheap professional carpet cleaning companies western sydney under $100", someone looking for paint colour ideas may search for "federation style paint colour ideas living room", and someone wanting for a holiday might search for "flight and hotel deals melbourne to fiji with kids october 2014".

All very plausible searches, and all very relevant if your business operates in those sectors.

If you're an Australian travel company selling Fiji holiday packages, for example, it would be fantastic to show the following tailored ad message for the search "flight and hotel deals melbourne to fiji with kids october 2014":

It's targeted, relevant, and caters for the specific requirements detailed in the user's search.

But scale this up to include all possible relevant locations, destinations, months, years, hopes, desires - and you soon realise you're looking a millions of potential permutations, and millions of potential ad messages, most of which only have a tiny chance of being shown.

So what do you do?

Data-driven long-tail expansion

Luckily, there is a way to reap the benefits of a long-tail Google AdWords strategy, while keeping your campaigns manageable and efficient. But it involves starting small, thinking ahead, and using real data to drive your campaign expansion. Using the cleaning industry as an example:

1. Start small - choose one generic service area keyword 'stem' to permutate with every location (e.g. 'cleaning')
2. Create tailored ads for each permutation
3. Let the ads run for about a month to collect data
4. After a month, run an ad group report to see which locations are searched for more than others
5. Using this data, create three separate lists based on search volume: high search volume locations, medium search volume locations, low search volume locations, no search volume locations
6. Armed with the knowledge of which locations are searched for more than others, build out your keywords to permutate other keyword stems (e.g. 'commercial cleaning' and 'patio cleaning')

However, as pointed out by Andy Holden on Google+, this 6-step process still leaves a lot to the imagination, such as:

• What match types should you use?
• How should you name your campaigns?

Cannibalisation is also another huge problem for long-tail Google AdWords campaigns, as described in Google AdWords Trap and Help! Google Are Showing the Wrong Ad.

So to help clarify how you might actually go about using this technique for your campaigns, below is a step-by-step walkthrough, using the cleaning industry as an example. To keep things simple, we're only using 3 basic campaigns for now, but feel free to scale it up as your see fit.

Step-by-Step Walkthrough

CAMPAIGN 1 (THEME)

Firstly, create a huge list of relevant themes in Excel, which describe your products or services (e.g. tile cleaning, steam cleaning, tile and grout cleaners etc). The number of themes will depend on the diversity of your product and service offering. Put all your themes in a campaign. This first campaign will contain non-location keywords such as 'carpet cleaning', 'steam cleaning companies', and 'professional cleaners'.

CAMPAIGN 2 (LOCATION)

Next, create a separate list of relevant locations (e.g. countries, towns, cities, districts, states, suburbs etc) which describe where your business operates or services. The number of relevant locations will depend on the extent of your geographical coverage. Combine you locations with the word 'cleaning', and put all your location keywords in a second campaign. This second campaign will contain keywords such as 'melbourne cleaning', 'bondi beach cleaning', and 'western sydney cleaning'.

COLLECT DATA

To create your third campaign, it is not just a case of permutating list 1 with list 2. If you have 200 themes and 20,000 locations, for example, this will result in 4 million keywords, which is impractical and will exceed Google's default keyword limits.

Instead, turn on CAMPAIGN 1 and CAMPAIGN 2 and allow them to collect search query data. If you look at the search queries, you'll notice that theme + location search queries will be matched to both your campaigns. This is fine - you'll be using these theme + location searches to build your third campaign.

CAMPAIGN 3 (THEME + LOCATION)

To build third campaign, you'll need to mine campaign 1 for popular locations searches, and campaign 2 for popular themed searches. Using techniques such as broad match generation and theme analysis, identify theme + location searches (e.g. bathroom grout cleaning bondi junction) which have sufficient search volume to justify being added to campaign 3.

• Any popular and relevant LOCATION searches you find being matched to your first campaign (e.g. bondi junction carpet cleaning), add as new keywords to campaign 3, and negatives to campaigns 1 and 2.

• Any popular and relevant THEME searches you find being matched to your second campaign (e.g. bondi junction carpet cleaning), again add new keywords to campaign 3, and negatives to campaigns 1 and 2.

The process of adding negatives to your first 2 campaigns will mean that generic keywords in campaigns 1 and 2 are guaranteed not to cannibalise long-tail keywords in campaign 3. As you continually expand your keywords, your funnelling negatives will prevent searches which could have been matched to a long-tail keyword in campaign 3 from being matched to a generic keyword in campaigns 1 and 2.

In other words, if someone searches for 'bondi junction carpet cleaning', you want this search to be matched to 'bondi junction carpet cleaning' in campaign 3, not 'carpet cleaning' in campaign 1 or 'bondi junction cleaning' in campaign 2. Campaigns 1 and 2 will therefore be solely used as a generator of new long-tail keywords for campaign 3, using data to guide your expansion. As keywords are generated, negatives are added to force long-tail traffic to your long-tail campaign.

• Over time, as your campaign 3 expands, the amount of location searches being matched to campaign 1 will decline. Most search queries being matched to campaign 1 will be non-location themed searches such as 'upholstery cleaning companies'.

• Similarly, your campaign 3 expands, the amount of themed searches being matched to campaign 2 will also decline. Most search queries being matched to campaign 2 will be non-themed location searches such as 'bondi junction cleaners'.

The aim is to get as much of your traffic as possible being sent to campaign 3. The higher the percentage of traffic which is sent to campaign 3, the more long-tail your campaigns.

Bear in mind that the examples above may need to be tweaked depending on the relevancy and scope of your business. Once you've mastered the basics, also feel free to scale it up or take funnelling negatives even further to meet the specific needs of your business.

Share your thoughts below

Feel free to post your questions and comments below. I'd love to hear from anyone who has tried this technique in their campaigns, or used an alternative keyword strategies to take advantage of the long-tail on a large scale.