Brand building in an internet-age is part art, part science, and part culture. Subscribe for idea's, insights and playbooks, developed by founders, operators and creators.
Nice work. See you in your inbox.
Our tech didn't like that. Can you try again?
Insights
Published:
August 23, 2024
How your customer engagement data can help you align brand building and customer acquisition
Dan Wilson
Chief Data Officer
Digital platforms (ie Facebook and Google) are great at optimising campaigns towards response and trying to push customers to conversion. This works great for low consideration, impact purchases. But what if you are selling something with a higher price point, or a longer-term subscription commitment?
People don’t tend to click on ads and immediately buy those products. Any business with a longer consideration cycle starts to hit the challenges of digital attribution, which is something I spoke about in my last blog post. The current algorithms, which optimise towards conversions, don’t care about the negative externalities of customers who don’t convert.
For example:
If one ad is shown to 100 people, 4 buy, 10 are quite interested but don’t buy, and 86 ignore it completely, that ad is favoured over another ad which is shown to 100 people; 2 buy, 50 are quite interested but don’t buy and 48 ignore it.
Most businesses would see more value in sacrificing 2 current sales for the 40 extra people who gain interest and they might be able to get to buy later, but the algorithms will not optimise that way (at least not by default, you can set up custom tracking to do this). And typically you will need a very different creative message to improve “quite interested” than you will to improve “buy now”.
💡 Top Tip: Knowing when to introduce the brand rather than sell the product can have a massive increase in overall performance. We have seen decreases in CAC of up to 40% by optimising the allocation of introductory vs conversion messages.
Join the fold
Plug into our data science lab to get deep insights into what drives revenue.
Nice work. See you in your inbox.
Our tech didn't like that. Can you try again?
People don’t tend to click on ads and immediately buy
People engage with these higher consideration brands differently, and your audience and website data will highlight this:
• Your ad platforms will show a higher conversion rate on higher ad frequencies. As people don’t tend to click through and buy expensive products from a brand they have never heard of before but they will be more likely to respond on a 2nd, 3rd, or 4th impression.
• Your Google Analytics will show higher repeat visit rates, and a higher conversion rate on repeat visitors than on first-time visitors (even when accounting for only first-time purchasers).
• You will likely see a difference in when you get the highest page views and engagements, and when you get the highest conversions (or best conversion rates).
This last point is really key, as it tells you how customers want to engage with the brand differently at different parts of their journey. Frequently we see this most prominently on weekly seasonality (how customers act differently on Monday – Thursday than they do on Friday – Sunday). We also see this on how they act differently on mobile and desktop devices, where we see mobile behaviours skew strongly towards research behaviours. Or we may simply see this change in behaviour from their length of time with the brand, where they show more engaged research for the first 2-3 site visits before being much more conversion focused on subsequent visits.
Even publicly available data can help us see this. Google trends gives us indexed search behaviours for any search term large enough. Take Peloton, if we look at when users search for “Peloton reviews” (a typical research behaviour) and when they search for “Peloton cost” (a good proxy for purchase intent) then we see they have an inverse relationship. Users are most likely to want to read reviews on Thursdays and Fridays, but most likely to start looking to purchase on Saturday, Sunday and Monday.
It is obvious that we should be pushing different content and different messages to customers who are in a research phase than customers who are looking to purchase here and now. And luckily the data you get from marketing platforms, web analytics and purchase platforms helps you to know which message is most relevant to customers at each stage.
If you look at the type of posts that Peloton put up on their social channels, you can see that they adapt their messaging to match this shift in behaviour:
Here we have been through the last hundred posts on Peloton channels and loosely categorized them into Community posts (talking about presenters, training tips, shoutouts to ambassadors etc.) and Product posts (pushing new product ranges, discount codes etc.). If you align this to our search behaviours we can see that when users are most in research modes Peloton massively prioritizes Community content, whereas when we go into weekend purchase behaviours there is a bigger proportion of Product posts.
It is also worth noting how few Peloton social posts actually try to talk about or sell products, which is also a point many other brands could learn from.
💡Caveat: This isn’t a scientific dataset; the classifications are subjective and on a small dataset. This data isn’t easy to get hold of and there are only so many images of young, smug people on expensive bikes that I can tolerate).
Here are a few steps you can take to understand these behaviours and start to tailor the content you deliver to the message the customer is most receptive to:
1. Use your data to understand how much research and engagement users need before they purchase:
• Take a look at Time Lag reports in GA, if you have a significant amount of users outside of 1 day conversion then there is a distinction between users researching and coming back later to purchase. (note, GA under estimates these time lags as it isn’t fully able to match different sessions to the same user.)
• Conv by frequency in Facebook analytics. If you seeing higher conversion on adsets and campaigns with higher user frequency then it is likely that you have a longer consideration cycle and gives a good idea of how many messages someone might need to see before they are ready to buy.
• Email open rate vs conversion, if your users convert better on second or third emails rather than the first email then this also suggests there is a longer consideration cycle.
2. Find out when your peak engagement and peak conversion behaviours happen:
Google analytics is good for this, but your search keyword data will also be really insightful. When do your purchase pages and purchase intent keywords peak compared to product pages and generic / category keywords.
• By device; commonly mobile skews towards research behaviour while desktop skews towards purchase behaviour.
• By day and daypart; commonly you will see different behaviours on work days and non-work days, and on daytime vs evenings.
• By user lifecycle; do your returning users behave differently to first time site arrivals?
• There may be some other factors that form a key part of this journey which are specific to your business or category (monthly cycles if linked to pay days)
3. Identify any key content or product affinities with these different behaviours:
• Which products react best to research vs conversion. Again your keywords and site analytics should give good insight into this. (We will have another thought piece on this topic later. Knowing which products to promote when is a massive opportunity for any business with a larger number of SKUs and there are some smart models you can deploy to sell the right product at the right time.)
• How can you amend content length and design for passive research vs active purchase. You will likely see a different in dwell times and page views per session when users are researching. You may need to make messages and journeys shorter and more efficient during purchase behaviours but longer and more engaging during research behaviours.
4. Amend your existing content delivery to match this cycle:
Once you know when and where these cycles exist, and which product and content matches best to each then you can amend your messaging to tell the right story at the right time. This can then be activated through a number of different places...
• Social is one of the easiest and most effective places to deploy. As Peloton do, sell the brand during research cycles and sell the product at purchase times.
• Email and SMS should always be personalised and adjusted to individual user segments. Highlighting the brand and the purpose during research cycles before promoting offers and products during purchase cycles can be really effective.
• If you are able to amend Website landing pages based on new vs returning customers and personalise for signed in users then this can be very powerful. Adjusting this to not just the aggregate trends but specific to the behaviour you see against and individual user or user segment can really scale this to market leading performance (but needs a bit more tech and data infrastructure).
5. Test, test, test:
If you did the good upfront analysis then your changes should work more often than they don’t. But humans are irrational and unpredictable so sometimes they won’t act how they are supposed to. Luckily all the same data sources that let us devise a new strategy will allow us to measure the outcomes and adapt for the future.
There is a lot of talk around choosing between data-driven marketing or creative brand building, but you shouldn’t have to choose. When done correctly proper analysis of customer engagement data should allow you to tell a more creative brand-building story, and measure the overall benefits that has on customer acquisition. We should be well past the point of idly chasing clicks and conversions.
If you'd like to know more about how our data science team can help drive growth for your brand, drop us an email at hello@charlieoscargroup.com. And for for more data science insights and breakdowns, make sure to sign-up to our newsletter - Cuppa - here.
