One of the biggest challenges in paid search marketing is measuring the impact of generic keywords on your sales.
We all believe generic keywords have a brand-building benefit, but how can you accurately measure this?
In a recent article, I talked about how many site owners are still having major issues with sales duplication within their online marketing, and are double and even triple counting their sales across multiple online channels.
The easiest solution for resolving this is simply to de-duplicate, and therefore assign all the benefit back to the last click. The challenge with this is that none of the marketing that influenced the sale receives any of the credit for doing so.
Within paid search marketing, generic keywords, such as ‘Holiday’ or ‘Insurance’ or ‘Plasma TV’, drive large volumes of traffic but have very poor conversion rates and return on investment.
Search marketing agencies often offset the high cost/poor return of generic keywords against the strong performance of brand keywords, meaning the overall average results of the search campaign are acceptable.
Break this down and report on keywords in isolation and the true picture is that generic keywords simply aren’t worth the money.
However, most search marketing agencies believe that generic search keywords raise brand awareness and that customers will eventually convert, probably via a brand keyword.
While I don’t dispute this, most do not have the facts to support this claim. In an industry that prides itself on being measurable and accountable, I don’t feel the approach of ‘inner belief’ is acceptable any longer.
Recently a number of search technology companies have introduced click-path or click-stream reports to give visibility into customer journeys.
These reports show individual customer journeys from their first search click to their last interaction before sale. This could be spread over 10, 30, even 60 days, and include two, five and even 10 search clicks before sale, depending on the typical sales cycle.
During my time at Warner Breaks, this report was invaluable. This was the first time I could truly measure the value of the generic keywords in terms of their influence on subsequent brand searches.
I could now clearly see, for example, that a customer first searched for ‘Short Breaks’, then came back a week later via the term ‘Hotels in Yorkshire’ and a week later via ‘Warner Breaks’.
Clearly in a last click model, the sale would have been attributed to ‘Warner Breaks’ – ignoring the influencers that had helped drive the sale.
Analysing the click-paths, we then built some simple attribution models to share a proportion of the benefit (or CPA) back to each of the search keywords that influenced the sale.
Whereas previously I thought I had reached a saturation point based upon an overall spend vs return model for the campaign, I could now clearly see opportunities to increase my portfolio and spend levels.
This visibility gave me, and senior management, the confidence to switch funds from other marketing channels and invest in search marketing.
In addition, this report gives clear visibility on which keywords are performing and which can be removed.
For example, I found that ‘Weekend Break’ performed considerably better than ‘Short Break’ in driving sales via subsequent brand term searches.
We also started to map out customer research and booking patterns, understanding at what time of day customers would most likely to be researching holidays and therefore when to up-weight our spend on generic keywords.
To whet your appetite, I will write soon about how we went about including display, affiliates and email marketing in this same report to give us visibility of customer paths across all online channels. Exciting stuff!
But for now, I will leave you to delve deeper into your search marketing stats and go further than just the last click.
Matthew Finch – view blog