BrightonSEO presentation slides: Beyond the last click
Apologies for the delay in getting these up on the blog, but below are the slides, video (thanks to Silicon Beach Training) and transcript from my recent BrightonSEO presentation:-
Beyond the last click: Finding hidden SEO value with Multi-Channel Funnels.
Nazilla previously wrote an excellent write up of the whole BrightonSEO event here on the Fresh Egg blog, as did @danbarker (What you may have learned at BrightonSEO) and Silicon Beach Training (BrightonSEO roundup – part 1, part 2).
For me it was a very enjoyable day – both as a spectator and speaker – and I enjoyed all the presentations, as well as the beer induced social event on Brighton pier in the evening. A big thanks to Kelvin Newman & Co. for all their hard work organising the event – already looking forward to the next one!
As this was a 20×20 style presentation (i.e. 20 slides with only 20 seconds for each!) the slides are more visual than text heavy. Therefore I’ve added the full transcript below which will help add a little context around the more abstract slides.
Beyond the last click: Finding hidden SEO value with Multi-Channel Funnels.
A Pecha Kucha (20×20) presentation by Dara Fitzgerald from Fresh Egg.
To add a little context around the need for Multi-Channel Funnels. In the modern digital landscape users interact with websites through an ever growing range of channels: search, social, email, affiliate and so on. Conversion journeys are truly multi-touch, with each campaign playing a role within the overall buying cycle.
The standard approach to attributing credit for conversions for many years has been the last click model. What this means is if a user visits a site multiple times through different campaigns or keywords, any conversions for that user will be credited to their last interaction with the site. So in this instance the introducing and influencing referrals get no credit for the resulting conversion.
Conversely with first click attribution, the introducing campaign gets the credit for any resulting conversions, regardless of how many additional visits take place prior to the conversion. So in this case, the introducer gets all the credit and the influencers and closer gets none. This model is often used to focus on the campaigns driving new visitors into the funnel.
To hijack a commonly used football analogy, relying solely on a last click attribution model may lead a manager to sack his midfielder for not scoring any goals. Despite creating countless opportunities he gets no credit as his name isn’t on the score-sheet. Similarly a first click attribution model may lead the manager to drop his striker for not creating any goals, despite finishing them.
The issue with any ‘single touch’ attribution model is that it will always be biased towards a particular section of the buying funnel without taking the overall picture into account. Removing keywords from the top of the funnel because they don’t convert will lead fewer visitors through the funnel. Removing keywords from the bottom will bring less of the new visitors back to convert.
So Google’s answer to the multi-channel attribution question is Multi-Channel Funnels. The overview page within the reports shows total conversions along with the additional assisted conversions metric. The Venn diagram shows the overlap between channels. So in this case we see 6.53% of the overall conversions involved both paid and organic search.
The path length report is useful for justifying the need for Multi-Channel Funnels as it shows the typical number of steps involved in conversion. So in this example, over 60% of the conversions involved more than 1 visit. Even more importantly in this example, over 70% of the conversion value was generated by paths with more than 1 step.
Similarly the time lag report shows the number of days between initial visit and conversion. In this example over 45% of the conversions (and over 50% of the conversion value) occurred after 1 day or more. Furthermore it’s interesting to see that nearly 25% of conversions occurred after more than 12 days.
The assisted conversions report is where it starts to get really useful for marketers. Alongside showing the standard last interaction conversion totals for each channel, this report also shows the number of times each channel assisted in a conversion where it didn’t get the last click credit. In this example, organic search assisted in over 1,800 conversions as well as converting over 2,200.
You can also view first click conversion data alongside last interaction to show the relationship between first and last click attribution and see whether a channel is more of an introducer or a closer. This cuts out assists and simply shows how many conversions were started and finished by each channel. Organic search in this case closed more conversions than it introduced.
The top conversion paths report shows the full paths taken to conversion during the sample period. In this case the top converting path involved two steps, both of which were direct visits. The default view will only include the basic channels tracked by GA, so direct, referrals, organic and paid search and a pre-defined social bucket. However these can be drilled down further using custom groupings.
Custom channel groupings allow much greater levels of granularity for focussed analysis. You can create as many as you like an examples could include splitting organic search into brand and non-brand keywords, or generic versus specific. You could also use custom groupings to split Google Shopping and Places out individually or affiliate activity and email.
Conversion segments are an incredibly powerful feature and are the Multi-Channel Funnels equivalent of advanced segments. These allow you to segment the conversion data so you can focus in on specific paths only. For example you could create a conversion segment that only included conversion where the first interaction was organic non-brand and the last interaction was organic brand.
A few important points to note about Multi-Channel Funnels. Firstly it works off a 30 day conversion window so any touch points prior to 30 days will not be included in the reports. If a campaign appears more than once in the middle of a path it will only count as 1 assist. And finally direct visits will overwrite an existing campaign and claim credit (unlike the standard GA reports).
So a few use cases. One of the main ones for SEOs will be to compare the relationship between brand and non-brand keywords. Generally speaking the generic terms you will be chasing will show a smaller conversion rate than specific brand terms. Nevertheless without the non-brand generics driving initial visits the resulting conversions may never happen.
So in this case you could create a conversion segment that only includes conversions where the first interaction was non-brand organic term and the last interaction was an organic brand term. In this case, in addition to the conversions reported as standard, there are an additional 79 which were assisted by non-brand keywords but converted by brand terms.
Another common use case would be to compare the relationship between organic and paid search. So similarly to before we would create a conversion segment but this time it will only include conversions where the first interaction was organic and the last was paid. This could be used to further justify the value of upper funnel organic key terms.
In order for Multi-Channel Funnels to work to its full potential and to get the most meaning from the data, it is imperative that you use consistent and robust campaign tagging. All campaigns from paid search to email marketing and affiliates should be tagged correctly with Google’s utm parameters to accurately track each channel.
In summary, most online marketers have been guilty in the past of knee jerky reactions and have been too fond of the ‘say what you see’ approach to conversion attribution. Not looking at the full picture means you can’t be sure of exactly what is happening and this can lead to very bad decisions. I urge you to start looking beyond the first and last click and start analysing the overall conversion journey as a whole.