Product Marketing Metrics: Present & Future
Think about a top chef’s shift for a second. They have to wash and chop vegetables, prepare sauces and desserts in advance, even put on an apron to help with the way too many orders — all while making sure the line chefs cook delicious food on time and dishes are being sent to the right tables.
So is the life of a product marketer, who needs to understand the technical implications of their products, map out user behaviors, and influence the customer journey by working alongside the product, support, and sales team.
Because of this, there’s an active debate on what product marketing metrics to follow and when to apply them, with the majority leaning towards the revenue-based ones like LTV and CAC. Although actionable, adoption and retention also play a big role in the well-being of a product. And company.
That’s why I’d like to share a set of simple, straightforward metrics that me and my team at Paymo use, as well as debate their future in the context of AI.
Sales
At the end of the day, a company does essentially two things: it creates products and sells them.
But what happens if you don’t have a clearly defined sales team with sales reps, account managers, and customer success reps, nor the capacity for one? You optimize for speed and track the following metrics:
- Average length of sale — for us, the moment somebody signed up for a trial until they become paying customers
- Average deal size — can vary for different customer segments, goes hand in hand with average length of sale
- Number of product demos/month
- Conversion rate from demo to paying customer
Website traffic
Usually, your customer’s first point of contact with your product is your website. This is your chance to test which pages, type of content, and call-to-actions better resonate with your customers — hence where to spend most of your efforts to attract and convert traffic. Look at:
- Monthly unique visitors
- Monthly number of trials
- Conversion from visitor to trial
- Conversion rate from trial to paying customer
Product/feature launches
Part of go-to-market strategies, product/feature launches are unique endeavours that bring awareness to a product, while adding value customers at the same time. Don’t be eager though, results won’t show up overnight. That’s why it’s best to set a timeframe for realistic outcomes (5% adoption in 30 days). The milestones you should have in sight are:
- Content views
- Social media shares
- Trials generated or up-selling revenue/new feature
- Product usage and adoption/new feature/timeframe
The future: automated metrics
These product marketing metrics are just the starting point, each product having its own set of goals depending on their industry and growth-stage.
Noticed something though?
I didn’t include the NPS (Net Promoter Score), mainly because it only tells you one side of the story — how likely is a customer willing to further recommend your product—without evaluating the actual reasons. Same goes for most sentiment analysis like social media likes or keywords trending.
Indeed, not all metrics are created equally. Yet the tide is changing.
To put it in Anda Gansca’s words,CEO at Knotch:
“Traditionally, brands have thought of people’s feelings as research, and people’s behavior as analytics.”
Not anymore, as qualitative metrics (like NPS) are evolving and becoming more granular thanks to machine learning.
As Dmitriy Pavlov, CEO at Stitched Insights, pointed out on LinkedIn:
We should totally be trying out new measures like how fully your product solves someone’s problem vs your competitors’ products (any indications of trend changes here can be super indicative especially with new feature releases) or tracking changes to how important top reasons for purchase/drop-off are to your customers along with competitor deltas (crucial to understand which levers are most important to your customers as they tend to change frequently, obviously depending on your vertical).
You could access this historical data with the help of a data scientist, who in the first phase can feed the algorithm already labeled variables to train it — like certain demographics for customer segmentation.
This is what’s commonly known as supervised learning.
What comes closer to artificial intelligence is unsupervised learning, where the algorithm can predict certain patterns without assistance, that would otherwise get unnoticed by a human mind. Like hidden customer clusters — to use the same customer segmentation example from above—that can be linked to a depedent variable, say one of the product marketing metrics already enlisted.
Read more about K-means clustering for customer segmentation here.
These two types of algorithms can help you build an open-source model, with more commercial “AI” tools to follow down this path like Stitched Insights or Knotch. Although I believe their applicability is more suited for enterprises.
Back to the kitchen
Until they become accessible to well, everyone, go for the product marketing metrics that allow you to make impactful decisions while keeping your hands in all the product-related matters.
Just like a top chef does with their kitchen 👩🍳👨🍳
If you’re serious about product marketing or know someone who’s just starting out on this career path:
- Clap this story 👏
- Share it with them 😊
