Having worked almost exclusively with retail websites for the last four years, I’ve spent a lot of time analysing data for different channels and trying to attribute value to specific marketing campaigns and projects.
Whilst doing this I’ve found a number of fairly obvious (only when you really think about it) potential threats to everyday attribution that I wanted to share.
Online marketing activity supporting offline sales
We recently saw this happen with a retail client we were working with who recorded significant growth in offline sales generated by the website. This success wasn’t being attributed to the SEO work we were doing, which was a big mistake on our part.
Attributing increases in offline sales to online marketing efforts is notoriously difficult, but regularly matching offline sales figures to increases in website traffic is the way that we are doing it at it at the moment.
For me, offer codes (not necessarily for a discount – could be for free next day delivery etc) represents the most effective way to track offline customers who have been influenced by your website and online marketing activity. There are obvious limitations with this, however running tests to get an indication of the percentage of people using the code would allow you to use a multiplier to get a better idea of how online is supporting offline sales.
The most common method for smaller retailers is simply to ask how the customer found the product, however this is often more difficult in commission-driven sales environments and people also often forget their first touch point. I would recommend adopting this principle anyway as it will help you to attribute value to specific channels as opposed to just the website.
Another slight win for this could be to promote ‘reserve in store’ functionality, which would help to attribute some offline sales to the original online impression.
Offline marketing activity supporting online sales
In the same way as online > offline, the effectiveness of offline activity is often under-attributed to the correct marketing campaigns.
In the past we’ve seen clients run offline adverts (promoting their website) without telling their online or ecommerce team, leading to false attribution to other brand-focused activity.
Attributing value from offline marketing to online is a bit easier than the other way around, as you can use vanity URLs to track how people came to the website and create segments to track how those people engaged with your website based on their landing page (which should only be accessible / promoted via the advert).
Another common way to track the value of offline marketing activity is to use QR codes alongside a vanity URL.
We had an example a few months ago where we were unaware that a client was running tube advertising for an event they were organising and the value was being attributed to the emails being sent out.
We then started using a QR code and a specific URL to help us attribute the value to the specific adverts.
PPC and display advertising not getting attribution from phone sales
This one is particularly common and can be detrimental to marketing campaigns when action is taken. It’s really common for people to look at PPC / display as a under-performing marketing channel when it’s actually generating a lot more sales / leads / revenue than you think.
We were working with a retailer recently who was spending a lot of money on PPC, which after having a brief look at their analytics I thought was a mistake. However, when I spoke to someone within the business I found out that they take at least 50% of their orders over the phone, which although they weren’t tracking, was obviously benefitting from their PPC efforts too. I've seen plenty of other people make this mistake too and then lose out by reducing their online ad spend.
In this scenario, I would recommend looking at call tracking (from someone like CallTracks or Infinity Tracking) to successfully attribute phone sales to specific channels and keywords.
Visitors choosing to buy products offline when browsing online (impacting conversion rate)
We work with a wine retailer who have offline stores as well as an ecommerce website. Something that we’ve found is that a large proportion of their visitors would prefer to find a product and then buy it from their local store, which inevitably influences their conversion rate.
We discovered this by monitoring buyer behaviour and again trying to attribute offline sales to increases in website traffic.
Subscription-based / on-going revenue not being reflected with bottom line figures
This is something that has been really frustrating for us, as it makes reporting a lot more difficult. Last year we built a finance extension for Magento, which is being used on one of our clients (and soon to be more of them), which only assigns 10% of the value of the order to a transaction being recorded.
This black hole also applies to businesses running a subscription model, as only the initial payment will be assigned to the transaction value.
There’s not really a way to counter this, as assigning value via goals or trying to over-ride the transaction value could be misleading, as there’s no guarantee that the user will pass the security checks or continue to subscribe for a set period of time. For reporting on deposits, I’d recommend taking an average for the number of people that complete the transactions and then using the full revenue of the products within a clearly labelled forecasted revenue field.
These are just a few examples of issues, there are plenty more that have annoyed me and loads more that I’ve not faced / thought of. If there are any other reporting black holes you can think of (eg: browser previews impacting data or PayPal revenue not being included within reports), please feel free to document them in the comments below.