Mysterious Marketing Metrics Mix
Which ones do I track and why ?
Within my current role, I typically go over a ton of reports on a daily basis. I have read a lot of articles and also wondered in the past how other CMOs make decisions when it comes to structuring the marketing metrics mix in their environment. It seems as much a mystery as any part of a marketing strategy. So, instead of building a view on what I think others do, I decided to share with you the approach I have taken in the last years.
No promise nor belief on this being a silver-bullet approach, yet I invite you to let me know what you feel through the comments below.
Raw business metrics are the ones I focus on every day, because for me those are the only ones that really matter. How is our revenue growing compared to last year and our latest projections, how is the profitability evolving and what amount of new offers, visitors, and buyers did we welcome?
Only after I fully understand the status of each country on these metrics do I look into the marketing metrics mix.
Start with your business metrics first.
Turnover — Profit — Offer — Visitors — Buyers
There is a set of ‘softer’ measures which I follow up on a weekly basis. The NPS (measured at different stages of the experience flow) and customer satisfaction metrics. We combine fixed satisfaction and NPS flows with qualitative surveys on specific sales that are either outperforming or lagging versus the forecast. Questions are crafted to get a better understanding on the why of the behaviour. This results in adaptations of commercial processes yet I do feel it is an often overlooked value marketing can bring.
High level marketing metrics are available to me through a set of standard dashboards (in Tableau, Mixpanel/Microstrategy, and, of course, good old Excel) that auto-update every night and can give me the overview of the metrics I am focused on at that moment , because the business results will always influence my focus.
Typically I start by looking into the sector mix (our sales are bundled into sections like FMCG, fashion, home decor, etc) to see the evolution in average turnover and visitors per sale, looking for trend information that I will compare with the market reports I received the week before. It is fine to have good results, but if the market had better ones, we missed opportunities or are missing an angle.
After I analyze how we are performing versus the market-trend I move on to the evolution of the customer base per country. I look into the customer segmentation mix (we use a custom RFM algorithm to cluster customers in segments) and how their average visit and purchase rates are evolving. Do I see older visitors growing in engagement on our platform or are we currently just banking on a lot of new visitors (due to the confinement, for example) of whom we have less view on their long-term engagement? Is the super (most loyal) segment still outperforming our forecasts and quickly becoming a dangerous 1% mix?
Based on the sector mix performance and the distribution of old versus new visitors’ behavior (versus last year and the week before), I try to answer the main questions on our current performance.
Can we fully explain what we saw yesterday and can that knowledge help us get a better result tomorrow ?
Only after answering that question will I move into the details of the performance marketing metrics — if needed.
If not needed, I just review our current product and project portfolio to see if it is focused on the right initiatives and drivers. Sometimes negative trends were already visible or predicted and we are working on a mitigation or booster plan that has its own specific dashboard.
“Needed” for me means that there is a deviation between the forecast and the actuals or a new forming trend that we didn’t yet predict or for which we have no project currently in place.
Now, imagine we saw a huge increment on an average Thursday (meaning no special weather conditions, no special marketing push, no outliers in offer mix) and we detected a lot of ‘churned’ visitors re-engaged with our platform; we’ll need to study where they came from.
Did our open and click rates grow on our emails, did they engage on organic & paid posts on social media, did we see spikes coming from referral sources? Based on these insights we will typically aim to grow the positive effects of one outperforming channel even more while doing a deep dive on the performance of the ones lagging in case we saw a huge drop.
“Results of the past are no guarantee for the future”
— all the investment info sheets
So, I only dig into the detailed metrics of outperforming or massively lagging channels. It is really easy to state that you should be on top of all them, and for a part, if your teams are big enough, chances are that at least someone is on top of each of them.
However, from a business supporting perspective, it is important to focus on what moves the P&L needle most. Focusing on too many marketing metrics can cause analysis paralysis and actually prevent your team from focusing on trends that matter.
One of the positive side effects of the approach we take is that we often shift budget from one channel to the other over the course of days. We shift our money to where the money is made. If each team has a set budget, and you only focus strategical changes based on big projects, they will focus on their platform-specific metrics and you will never be able to fully bank on outliers as a company.
So yes we look at the cost of exposure (cost of 1000 views), cost of the actions (typically signing up on our platform) and track the success of purchase-automated retargeting (cost of buy action) for each channel in their own reporting as well as with an attribution model based on our own database events. We do look at the value (customer lifetime value or changes to it) of each action to see if it makes sense from an ROI perspective.
But mainly we focus on the business results and look how we can positively support those in the long term.
We focus on the evolution of the existing customer base through customer lifetime value and the triggers that change it while monitoring the evolution of ‘new’ traffic and conversion, ultimately leading to focuses on both
CLTV and COBA.
That does not involve setting up click bait SEO blog posts, does not involve creating multi-month projects to optimize look-alike value-based audiences till infinity for a specific platform, nor do we overextend our amount of A/B tests on mail title optimizations. We focus on clear data trends and then optimize the channel setup to a level that makes sense in the context of today and tomorrow. We A/B test every change we make to avoid being led by our own bias instead of focusing on what the data and thus the real customers tell us.
I haven’t focused on branding, PR, and content marketing in the article above, because I do feel that those are harder to correctly measure. They should support the long-term vision of the company rather than be judged on short term effects, in my humble opinion.
I am convinced that sharing an honest view on the beliefs of the company and the value it — or its product — brings, combined with a relatable core message and consistent tone of voice is as valuable as the whole performance marketing setup, especially for those who are in it for the long run.
