From measuring traffic to driving company strategy: The data journey of a (fictional) mattress company

TL;DR: Your data needs evolve with your business needs. Here’s how, based on the hypothetical data journey of a mattress company.

Over the past 10 years, Snowplow has helped thousands of companies drive value with behavioral data. During this time, we’ve noticed that the data challenges businesses face at each stage of the data maturity journey are often comparable and can be mapped out predictably onto a data maturity model. The needs and ambitions a company has when they are just starting out will look very different to the needs and ambitions of those at the other end of the spectrum. Awareness of this trajectory allows businesses to predict common challenges and plan for their solutions, helping to navigate the journey towards data maturity.

The journey ahead speaks specifically to behavioral data, rather than transactional or demographic data. Read this blog to learn more about the difference between behavioral data and other types of data.

Matratze Platz is a (fictional) Frankfurt-based mattress shop, run by one-man-band, Stefan. This article follows Stefan through Matratze Platz’s hypothetical journey to data maturity, as modeled on the experiences of our customers and industry experts. We see how Stefan uses behavioral data to advance specific use cases over time in order to drive growth, and how he overcomes some common bumps along the way.

Level 1 – Data Aware

Business needs: Understand basic usage

At this stage Stefan’s needs are not particularly focused on data. With the view that customers should try-before-they-buy by coming into the store, data on Stefan’s online traffic is of little interest to him as his focus is on in-person sales. He sees internet mattress shops as marketing companies that happen to sell mattresses, and is loyal to his more personal, experiential value proposition. 

At the twist of an arm, Stefan might decide to find out whether people are actually visiting his website. Having a storefront, he may want to know the geography of these visitors, but his needs at this stage will not extend much further. As a one-man team, Stefan simply does not have the time or resource to be collecting extensive behavioral data and asking sophisticated questions of it.

Level 2 – Data Capable

Business needs: Acquire new users with targeted marketing

Following this endeavor, business is going well and Matratze Platz is growing. As the company grows, its needs begin to change. Currently Stefan only sells mattresses in Frankfurt, but what if Stefan wants to grow his company to also supply Hamburg, with a view to expanding across all of Germany? 

At this stage Stefan moves beyond only collecting data, to actually acting on it in order to show everyone in Frankfurt and Hamburg an advert with his latest offer. To do so, Stefan needs to put some basic criteria on his targeting – he knows the age range of people he wants to target, and with the help of a newly hired marketer is able to locate the destinations that his target audience frequents, and so places an advert. 

The success of this campaign prompts Stefan to consider selling bed sheets in addition to mattresses, which raises a number of questions:

• Are people going to be interested in buying bed sheets from Matratze Platz?
• Do people buy bed sheets when they buy mattresses?
• Are people likely to come to Matratze Platz just to buy bed sheets?

The questions Stefan is asking are getting more sophisticated. At this point he also knows anecdotally (not from data) that people don’t tend to repeat purchase mattresses, but that people do buy bed sheets when they buy a mattress, and then repeat purchase them quite frequently. 

Level 3 – Data Adept

Business needs: Retain users through deep understanding

Matratze Platz is now expanding beyond Germany into Austria, and Stefan wants to know whether buying habits differ between the two. To find out, Stefan hires his first data practitioner. If data shows that buying habits do differ, then Stefan needs to know how best to act on that information. 

“How should we bundle our products – bed sheets, bed frames, mattresses – in order to create repeat customers?”

At this point, the marketing team are running regular campaigns across social media and web, taking a scatterbomb approach in order to tell as many people as possible about Matratze Platz’s offers. Whilst this increases reach far beyond the capacity Matratze Platz had in earlier stages of its data journey, there is no real sense of whether the audience being reached are actually in the market for mattresses. Factors like whether someone has recently bought a mattress, or whether they are moving house, will determine whether someone is likely to buy a mattress (or not). Matratze Platz’s current approach to targeting customers is therefore not particularly sophisticated.

Targeting people who are not going to buy a mattress is not an effective way of deploying the marketing budget. Likewise, it makes little sense to target digital ads to people who have already decided that they are going to buy a mattress – for example in the few days leading up to their purchase. However, two months before buying a mattress, ads could sway someone a lot. More intelligent targeting is needed. The data team starts to become a bottleneck for providing data for this and other increasingly sophisticated requests from teams across the business:

• Buying and Merchandising wants to start using data to forecast buying trends
• The Search team wants to know whether people are buying products after searching for them, and is looking to start personalizing search results in order to boost engagement.
• The Product team wants to use detailed search and engagement metrics to inform their product development.   

And so on. This marks an inflection point in the journey. A lot of people from the business are building their own reports. There’s a lot of demand for data-driven insight. The questions being asked are highly specific to the business model and the intricacies of the business. The fact that Matratze Platz has two or three different products, selling mattresses and bed sheets, makes for vastly different and highly specific customer journeys. Sheets are more often an impulse buy with a short purchase cycle, whereas mattresses tend to be planned purchases once in a decade. This creates nuanced requests of data.

Matratze Platz is now powering a lot of different use cases with data, from understanding and predicting customer usage, to targeting customers based on propensity to buy. Additionally, teams across the business are increasingly measuring the impact that they are driving, using data.

What does the inflection point represent? 

At the very start of Stefan’s journey, when he was operating alone, if he had been suggested an end-to-end point solution specifically for mattress companies, modeled on all the specific assumptions for mattress companies (expensive, one-off, local purchasing), he would probably have been quite happy. Stefan would also have been happy with a tool like Google Analytics, because it collects, models, stores and analyzes his data for him. This is useful when you just want to know what’s going on. 

However, at the point where the data team is fighting off multiple requests from around the business, which are actually quite complex, they are going to have to keep buying a point solution every time there is a new question, especially when the business model is so unique. These solutions are good for certain industries, but not for all, which creates a lot of frustration, and a system that may look like this: 

At some point in the last few years there was some requirement for each of these point solutions. But as time went on, more and more custom requirements came up. For example:

• Are we getting repeat custom from bedsheet purchasers?
• Do bedsheet purchasers start buying bed sheets from us after buying their first mattress?

The questions get really specific. The tools are all working from the warehouse, but they’re using bi-product data; this is data collected by different SaaS tools, which data engineers have to wrangle and piece together in order to answer questions. It’s an inefficient, reverse-engineered process, impeding data teams from driving greater sophistication going to the warehouse and being modeled in very similar ways. 

There comes an inflection point in Stefan’s journey where the sophistication of his concerns warrants a switch from using multiple point solutions to an investment in the modern data stack. At this point, switching to a solution that can adapt to the evolving requirements of the business is very sensible. Before this the associated investment required in both tooling and expertise is likely too large to justify, but when this inflection point is reached, the switch becomes very valuable.

Level 4 – Data Informed

Business needs: Personalize the user experience

Where does Stefan go from here? He now has the flexibility to go further, and is starting to get far more ambitious than before. Matratze Platz now has lots of people browsing online, and store clerks are also spending lots of time with customers. Stefan decides to equip every store clerk with an iPad, so that they can see, with the customer’s consent, everything the customer has done online. Mattress fitting and order processing can then all be done via the iPad and tracked alongside web data during the research phase in the customer journey, meaning that all the data is in one place, connected as one customer journey.

The key theme of this stage is evolving sophistication, and taking business to the next level. Matratze Platz is now a very large company, spending millions on advertising, and on marketing in general; departments where you cannot directly measure the impact of brand efforts on revenue. But Matratze Platz really needs to, in order to justify their spending on these business functions. 

Stefan toyed with attribution modeling years ago using GA, but it’s cookie cutter attribution models are a far cry from the accuracy Matratze Platz needs to attribute in accordance with it’s highly specific business model. At this point custom marketing attribution is essential.

Matratze Platz now has data resource within individual teams alongside its central data function. This means that it can start thinking about recommending products digitally via the website. For example, if a user has toggled between a double and a king size, this could be interpreted as a sign that this user might be willing to spend more. Recommending high-end bed frames to this user might therefore be of value. 

Being able to recommend products based on what we can assume from a user’s behavior is important. A similar recommendation to a customer that has indicated that they are shopping on a low budget might deter them from purchasing from Matratze Platz altogether. 

Similarly, we learned from Matratze Platz’s expansion into both Germany and Austria that buyer habits differ in each place – being able to use that customer understanding to recommend products differently to each of them is important. 

Level 5 – Data Pioneer

Business needs: Company strategy driven by behavioral data

Now Matratze Platz is a truly global enterprise, selling mattresses across the world and operating multiple data teams – it is now very clear how different business needs are compared to where it was at the start of it’s journey. At this point, Matratze Platz has dozens of teams, all starving for data. These teams need to be enabled without the central data function becoming a bottleneck.

The central data function needs to socialize data in a compliant, well-governed way. To do this they need to ensure that taxonomies are consistent across teams globally. If the Australian branch of the company is giving IDs certain formats, it is important that other branches are doing the same. That way, the New York office can easily use data from the Australian office, and vice versa – data that is generated by one team can easily be reused by another (in a way that is compliant). This applies cross-functionally as well as geographically. The aim is to give as much of the business as much data as possible, whilst also respecting user privacy and not abusing their trust.

How will Matratze Platz know that it has reached the stage of data pioneer? You know that Netflix, Airbnb and other tech giants have reached this stage because of the contributions they make to the broader data community – firmly establishing themselves as data leaders. 

At the stage where every team needs data in a compliant, well-governed manner, structure starts to take the form of the data mesh. Control over how data is generated and consumed is decentralized, but through a common understanding that ensures proper governance is maintained. This is a new framework emerging for businesses further along the journey looking to meet their data needs. 

If you suggested a data mesh structure to Stefan at the first stage of his journey, where he was just trying to figure out what was going on with his website, he wouldn’t know what to do with it. He has no need to explain how the data is governed to his team of one.

Driving value with the Modern Data Stack

The data journey of Matratze Platz is fictional; it’s purely based on our experience over the past 10 years of helping organizations to drive value with behavioral data. Whilst the story of Matratze Platz is hypothetical, it does demonstrate a trajectory common to companies looking ‘to do more’ with behavioral data. 

This trajectory starts with asking simple, foundational questions about their audience and products – questions that can be answered with packaged analytics solutions. As these questions become more complex, however, different solutions are needed.

At a certain point in the journey, we believe the only viable solution is the modern data stack. Powered by behavioral data captured by Snowplow BDP, a modern data stack liberates you to ask increasingly sophisticated questions and adapt them as your business changes. This means you’re free to drive significant commercial value with behavioral data, and deliver results such as a 100% in subscribers or a 1,400% boost in paid ad revenue. In short, by gaining a deeper understanding of your customers, you’re able to secure the future of your company.