Keeping up (data) appearances

Keeping things in order, making things look tidy, ironing your super creased shirt you’ve not worn since February 2020 are all things we do in everyday life. So why should we treat our data any differently?

Having moved roles to focus purely on data, naturally, I’ve never spent so much time thinking about it. I’ve previously written about being data-informed and that was when my interest in ‘data’ piqued. Today I’m sharing some thoughts on being data-centric. There’s a few elements I’m exploring, so hopefully it’ll be useful, no matter the role you’re playing in your company. In this post I’ll be focusing on:

1. Five whys for collecting/using data
2. Obsess over your data, not your models
3. Data utility – make it meaningful

Mo Data Mo Problems

There’s always something gathering data; Apple tracking your steps, Spotify taking note of what you jam along to (or don’t) for 30 seconds or more, Nest working out your home/away patterns, and Starling seeing how much you’ve spent on coffees…

These examples all lead to useful outcomes, whether that’s a behaviour change — like Chanade, you achieved a record in step today, or learning to play you more of the music you love in your daily mix. Whilst these are great, there’s a lot of stuff that happens where data is gathered but it just sits gathering dust, a bit like my suitcase at the moment. Many teams are quick to collect data, but slow to do something with it, especially something meaningful. Yet the other hand, you have teams setup to make data available and usable, so teams can create something great, whether it’s to support business teams or customers to make their next experience better than ever, because it’s relevant to them at that moment in time and that job they’ve come to do.

Some other examples of using data for the good of the customers include Zara. Zara have been on an impressive journey over the years, and customers’ are at the centre. Every day, store team members from the 2,259 stores in 96 countries feedback customer habits, which design teams then pick up to make improvements, this includes high-selling items as well as returns — this enables Zara to move quickly from a design and production perspective, for example this can lead to discontinuing a style, or altering the sizing.

Another fashion powerhouse doing good is ASOS. There’s a couple of things they’re doing that I found useful by leveraging return data, customer reviews and a little feature called ‘your details’. Super helpful when buying digitally, and in my experience has always been pretty accurate.

It wasn’t that long ago I watched it, but Moneyball is a great film showing the application outside of ‘digital’ products. In Moneyball there’s baseball team that hires an analyst, he focuses on a set of data features to inform who they play and who they sign/swap. This data is used to predict the winning team and long story short, there was success over the ‘always done it this way, this data was won’t work’ behaviour from the scouting team. It’s common behaviour now, and goes to show how far and wide data goes into predicting things across many industries.

Five whys for using data

If your team can’t answer why 5x on the data they’re wanting to collecting/use, then they shouldn’t be collecting it at all. To make this process more structured you could create some clear principles for your teams to stand by – no matter where they live in the company. This could avoid the circles that often people go around in when dealing with data privacy teams. Principles should be in plain English and meaningful. They should always be about enriching the customers’ experience when they interact with your products and services.

I’ve been thinking about this recently, and the agile manifesto feels like a good way to write these. So here’s my thoughts on principles for collecting data.

• Make it part of the experience over creating more effort for your customers
• Focus on the minimum viable data to get started over trying to think of everything that you might want
• Demonstrate the data's use in prototypes over providing a list
• Bring your privacy/security people on a journey over asking for permission all the time

Relating back to these four principles, team members put the customer at the centre and ask themselves a few things:

All of this will make your life easier when you’re working with the likes of data privacy, or regulatory to back up why you’re collecting/using certain things, as well as ensuring it’s ethical — this should be part of your considerations, always. Would you be happy if this data was being used to do X?

When it comes to collecting data, we can take a lean approach. There’s some great examples where companies are collecting the minimal viable amount of data to get you signed up to their product or service. Then as you keep going back for more they progressively build up what they would like to know to make using their product/service better. Adobe does this well with a focus on account recovery to begin. You complete the bare minimum and you’re in, then on going in again they tease out a little more. It’s the MVD to get going. It’s always good to start here, show value and scale over time, otherwise you’ll spend weeks getting teams to approve everything, weeks figuring out what to do with it all and less time on creating small value quickly for your customers. Focus on a small bunch of use cases, less is more. Always.

Obsess over your data, not your models

A lot of techies get obsessed with models, rather than the data. The problem is, that the data is what unlocks the value, the model is a means to an end.

‘data eats models for breakfast’… in the same way that ‘culture eats strategy for breakfast’ ??

When I think of it, it’s a little bit like making a sandwich. Let’s assume the model is the bread. It’s baked just nicely, always room for improvement but overall it’s a lovely loaf, the Warburton family would be proud. Then the data, let’s assume that’s the filling. Personally for me, if the cheddar is anything less than extra mature I’m not okay.

So, you go to the fridge (data storage like Google Cloud Platform) and it’s literally all over the place, you can’t find the cheese, the butters right at the back, you’ve got some onion chutney but it’s out of date. It might sound a little bit like how your company stores its data. The importance of data storage is often forgotten. This is fundamental to a high-performing data team, that create value quickly and build on that iteratively over time. How you store data really does matter. Cloud providers have decoupled storage and processing, and done a ton more to give people the tools to take data storage seriously. It’s a craft, appreciate these people and invest in them.

When company’s see the value in the journey data goes on, investment in modelling it is not a problem, but if it’s stored like ‘the man drawer’, you’re making your teams lives harder, and creating unnecessary effort. Be the Aldi middle isle of data storage — you come in for two things, and you leave with a welcome mat for your porch, a candle and a new set of pans. Organised data can lead to lightbulb moments for teams, and create innovative products.

Finally, in practice, your models built and you’ve shipped it, brilliant — but that’s not the end. It’s either doing great, it isn’t performing as well as an existing one, or it’s brand new, new territory and it’s not performing as well as everyone thought it would. Many people assume that the model is the problem . Hello, have you seen the data situation?! Take your data seriously.

Data utility – make it meaningful

Too often companies interact with me in the same way that they do with Pedro Pony, Rebecca Rabbit, Grandad Dog and Zoe Zebra (Peppa Pig fans know). It’s like they assume we all have the same job to do with the same thing, at the same time, in the same channel.

There’s a few things that have/haven’t made me think awesome, I like your style, please collect and use my data some more.

Virgin Wines – I purchased one case, now I’m targeted weekly with a 20% off email and/or text message, which I’ve never taken them up on, but I haven’t been bothered to unsubscribe. Feedback loop, where’s the change in approach? ??

Majestic – I go in store and buy 6 bottles, got to love the mix six. I’m a member, yet I get nothing. Not a thing. ?????♀?

Mango – I was browsing having seen an add for a gilet that’s totally a bit of me. Out of interest, I used the size feature. It helps you figure out what size you need. Part of this was pre-filled, my age, height, weight – useful. They’d tapped into data through iOS – I was fine with this, it supported a job to do and saved me time. ??

Classic example, but not in the usual light I’m afraid. Netflix & Prime. Sorry but I’ve watched 100s of shows, why am I still seeing them in all their glory? Why aren’t they in a little area like you’ve watched all these things or at least give me a sign that I’ve watched them, please. ?????♀?

Olay has a skin advisor that uses artificial intelligence, a selfie and some questions to determine the right products for you. It saves time, it increases the chances of it being just right and it’s a little fun too. ???♀?

Some thoughts on what you can do next ??

We’re in a place where teams will be on a mission to make your lives easier, or enrich your customers experience through the use artificial intelligence. You were quite happy to go to the shop in your self-driving car, or watch that show Netflix recommended, or breath because Apple Watch told you to, so be on-board and look to support opportunities where this technology can make your day to day easier, or your customers’ experience 10x better. Data is changing how we think, how we approach things and how we serve our customers better. Don’t get left behind.

• What you collect is important, but you need to understand why you’re collecting it. If you can answer why 5x, then when you come to document and discuss with privacy teams, you’ll have a solid vision.
• Data and it’s application unlocks value. It enables differentiation in a competitive landscape – or in Olay’s case your next foundation.
• Focusing on data cleanliness, availability and usability, not just shiny models, is critical to the value you’ll reap from using data, start as you mean to go on.

Being on the ball with customers and the jobs they trying to do leads to staying relevant, or in new entrants cases, can lead to disrupting entire industries. Data is critical to this, especially in this pricing example that follows.

Netflix were obsessed with knowing what customers were doing, whereas Blockbuster were fixated on 15% of revenue coming from late fees (70% of profit), to stop and see what was happening. Company’s need to think about data more widely, not just tracking what customers do on your website and build products that enable you to be the leader in your industry, and pivot through business model innovation to stay relevant. Sometimes short-term loss for long-term gain is necessary to have a breakthrough.