Many of your customers are e-commerce companies, what is the #1 piece of advice you give them if they’re starting with attribution during this season, or just wanting to adjust the way they work with your attribution data?
First of all, which is obvious but it’s easy to forget, the holiday season is only once per year so you should really make sure that all of the data gathering is in place. Meaning that, if you have Google Analytics in place you need to validate that, you need to validate your tracking systems to make sure they’re tracking correctly, and make sure you’re getting in all of the cost data in a standardized way. Because if you haven’t done so you might not be able to look at all the data points.
So, going into this season you should be thinking about if your setup is done correctly, and if your marketing intelligence and measurement setup correctly before the holiday season really starts to get going.
How do people go about making sure their Google Analytics is validated?
There are different ways. First of all, just compare whatever you’re seeing in Google Analytics in terms of conversions and revenue to what you see in your actual backed, or your BI tool. So talk to your BI analysts, talk to your software engineers, show them the numbers you have in Google Analytics and if the gap is much more than 10% then you should start to investigate.
That’s the first step, and then also try to understand the deduplication in Google Analytics. So if you look into Facebook and then you look into Google Ads and you sum up the conversions that you see in each platform you shouldn’t wonder why that’s more than you have in Google Analytics. And then try to find a way to manage those differences and how to handle the duplicated data.
What advice would you give for the type of model to use?
Depends on your marketing mix. So if you’re only advertising on Google Ads or Facebook Ads then you can just look into only the ad platforms. But, then the question is why are you not trying out other channels? Then it’s also about understanding how long your customer journey is, and how many touch points you have. If 80% of your customer journeys have only one touch point, you don’t really have an attribution problem - I wouldn’t waste too much time discussing whether first-click or last-click on your one touch point journeys is a better model.
But if 30-40% of your customer journeys have 2 or more touch points then it might start to make sense to think about something like linear attribution or u-shaped attribution where you at least give a fraction of the credit to all touch points just to see it being attributed more fairly.
I think those are the things to consider [marketing channel mix + customer journey touch points].
One way you can start to figure out how many touch points you have is to simply look into Google Analytics and you can get an estimation of that - obviously views from social channels are not being included, but you could get a good informed guess from GA.
Also, going into holiday season, you should make sure that you reach an agreement in the company before you go into the season - know how you’re going to evaluate the success of the season in terms of marketing. And don’t start doing this once all the spend has happened.
Now, I don’t know many e-commerce companies that don’t advertise on social platforms during this time, and Facebook is obviously the big one. They recently released that they’re changing their attribution window from 28-days to a 7-day window. What does this mean for e-coms during this season and how should they make sure they’re properly measuring the impact of social?
So there has been a lot of talk about this potentially decreasing the efficiency of Facebook campaigns. The short version is, this is not true; this is actually kind of just changing the attribution model that Facebook is applying to their own data and campaigns. It can, in the end, lead to some decrease the efficiency, but it’s not the direct thing that will be happening. And then, marketers who are worried about this should ask themselves, why did I allow Facebook to credit itself with the attribution values in the first place anyhow?
You should always have an independent system, either Google Analytics or Adtriba, or just some other way to not be fully reliant on Facebook data.
This also comes back to understanding how Facebook and other social channels work, obviously they don’t just work by people clicking on the ad, but the more prominent impact for social can also be the branding impact. If you see an ad on your Instagram or Facebook feed then you don’t necessarily click on it, but still your neurons might have fired and you remember, for example, the orange Funnel logo whenever you think of aggregating data and you just jump onto the Funnel.io website directly. So accounting for this view impact is also something that you probably should think about starting to do during the season, and if not during this season it’s a good thing to think about doing since it’s a step to more marketing maturity.
So this change that Facebook is making is not as dramatic as people are thinking.
Any final advice for e-commerce companies for the holiday season?
Just be sure to track the correct data, to get the cost data in place and to make sure you’re getting that in a standardized way.
The money spent on marketing isn’t only for getting the highest Return on Ad Spend, the lowest CPO, or CPA, but it’s also for an organization to learn and to evolve. You’re getting so much data back from the channels and vendors during this time, and if you don’t use these insights you’re basically wasting money. You might have a good campaign efficiency, you might have a good holiday season, but in the end you’re not getting the most out of your spend. So the learning is such and important part that you just need to make sure to have the tracking validated and other data collection setup in place.
And also take some time in January and February to do a retrospective and analysis of the hypothesis you had going into the holiday season of which channels will work and which channels don’t. If you find out that your newest campaign or channel has really worked well that might be an indication that you should be more aggressive and innovating with new channels - maybe even doing that the next holiday season even if it felt really scary.