JD Engelbrecht, MD at Everlytic, says the reality is that most of us are struggling to access data, make basic sense of it and activate it efficiently. As businesses and consumers, we are accumulating data faster than ever before.

Today, we have progressed far beyond the human ability to act on the firehose of data that flows through all organisations. Fortunately, sophisticated technology creates insight from chaos – and presents the potential to establish, maintain and grow individual and unique concurrent relationships with millions of consumers that simulate human intelligence. This is the dream. The promise. The possibility. For marketers and business leaders struggling for growth, it hints at utopia. 

Yet this dream has been sold (often oversold) at marketing tech conferences for years by people with ‘evangelist’ or ‘ninja’ in their titles, and we have been bombarded with jargon, buzzwords and complicated systems that seem to unnecessarily complicate a complicated matter.

Admittedly, we make it harder than it needs to be. Most often we get in our own way. Generally, the executive mandate isn’t wide or deep enough to enable those who have the ability to execute – and top brass does not kick down enough doors to enable execution teams to achieve what they need to.

Often, we buy tech (most frequently from the vendors who oversell the dream) that need consultants to operate and multi-year implementation schedules. The time to value is generally too long and the capital expenditure to get the ‘solution’ up and running is most frequently simply too much. High cost plus slow deployment equals frustrated and annoyed execs. This is particularly amplified when implementation runs across project sponsors’ tenures and new people inherit the programme.

So, where to from here? We need to reframe and understand where the challenge – and the solution – really lies. I am a firm believer in starting with the end in mind, but you can’t skip the things in between. There are foundations to build, lessons to learn, and there are corporate legacies to navigate. Your strategy needs to be primed and groundwork needs to be done.

Internal coordination and resource alignment

Primarily, we make the common mistake of believing that a data and automation strategy is a tech problem to solve. It is not. It is an organisational coordination and resourcing challenge.

So, if you think you are going to buy tech to solve this challenge, you are going to have a bad time and you are going to hate life for a solid stretch. Technology is absolutely the ultimate enabler – you cannot do this without it, but you cannot lead with it. In short, do not buy or build tech until you know exactly what you have and have mapped exactly what you want to achieve.

Tech can never fix a broken strategy, crumbling foundations or lack of experience. What it does do really well is make mistakes really expensive (and high profile). Importantly, once we get going, it becomes evident (but not always noticed) that we did not give the people we hold accountable for the outcome a strong enough mandate to get things done. Often we do not give the task to a strong-enough person, because you need some gravitas, guts, stamina and a thick skin to get this done in a large organisation.

Finally, you need to know what you aim to achieve. What are your organisational goals here? Goals disaggregate to objectives that are achieved through value workflows and conversations across the organisation. For example: do you want a customer to buy something? Map your sales conversation and funnel and fire it off at the right time, in context, personalised and relevant to a prospect. Now do this on an individual basis, at millions-scale.

Is your business model what you think it is? 

To effectively define your objectives and goals for a data strategy, you need to interrogate your business model, and its core purpose. So, what is your core business? This is a trick question. Yes, you might sell goods or services, but you are also a data, attention and trust business.

In the data – rather, insight, attention, and trust market – we all compete for a limited supply of these valuable consumer resources, and we compete against a near-infinite demand from businesses that overwhelm our consumers. With this in mind, we must responsibly leverage the data at our disposal, grab attention from anything that takes focus away from our solutions – and build a bank of trust that facilitates our relationship with consumers around our core business. Data, trust and attention are inherently linked – and form the combined platform to execute your marketing automation strategy from.

Building the foundations

With this said, we must start building some foundations to compete in this insight, trust, and attention economy.

Whilst the execution through automation is sexier, getting your data footing in place will determine your success. Let us take a quick look at the types of data at your disposal:

  • Descriptive: these are things you could probably determine if you looked at a user in person (gender, location, age, etc.)
  • Transactional: these are things you log as a result of transactional engagement (products bought, basket size, the timing of transactions, etc.)
  • Contextual: these are things you observe based on users’ engagement with your platform (position in the conversion funnel, device patterns, engagement patterns, etc.)
  • Inferred: these are produced as the results of running data through augmentation models to provide advanced insights (conversion probability, engagement scoring, predictive modelling, preferences, clustering, etc.)

You will find that 80% of your success comes from insights that are quite easy to obtain. In other words, you do not need expensive data science teams to unlock your first significant tranche of value.

That said, the challenge that we face is that we have created irrational system complexity with hundreds of federated solutions holding its data hostage in different formats, on different schemas and where only functional and practical integration exists to execute an operational task. Once we have made this data accessible, we can get to execution, right? But getting this data timeously, accurately and consistently is so very often where organisations fall short.

Creating a universal, accessible, data pool

At this stage, we must surface this data and create a universal data pool that is consistent, understandable, accurate and easily accessible via an integration layer that will expose the holistic data (which can be augmented first of course) to other systems for operational efficiencies – and to develop insight using various techniques from simple reporting, to advanced analytics, to data science and machine learning.

Importantly, you can surface the data and provide it for use without having to overhaul your entire technology platform. There is so much inertia in large organisations’ tech clusters that skipping the queue by integrating at a data layer will get you to value significantly quicker and with less pain.

This is a tough feat in itself, but until this task has been completed, you cannot truly unlock the value you need. That said, keep in mind that perfection is the enemy of progress and that you need only have the base data available to start creating value. But your end game requires a holistic view of a customer and the data that describes them, and their relationship with you and your products.

Getting your team in place and on board 

Cross-divisional participation can work if the overarching execs are closely aligned, and there is no conflict in their KPIs and incentive schemes. However, it adds another layer of complexity and drag to something that will produce enough pain without us adding more.

I mentioned that we do not give a wide, deep and strong enough mandate to a strong enough candidates to get things done. You need to appoint a strong person who has matrix competence that can own and add value to the entire data value chain. If you are serious about being successful, you must realise that this is a meaty thing that needs full-time resources allocated to it, whether they sit across your existing divisions or are concentrated into a centre of excellence. 

You will have an unusual, composite team that includes one or more integration developers, data developers, data scientists, data analysts and marketing tech specialists – and if you are serious about conversion optimisation and agility in terms of creating collateral, you need access to a designer and copywriter too. 

Then, the head of this band of misfits needs to be a good story-teller; be able to hold conversations with and contribute meaningfully to each dimension of their team; understand the strategic and commercial realities of each division of the organisation to bring them value; have a thick skin, and be fully assimilated into the organisation. 

Once you have your internal resources and team in place, make sure to take care of your unicorn and be tolerant of failure, as long as the overall trajectory is upwards. You will need to accept the fact that there will be failure and it is important to create a safe environment for failure and throw-away work. You cannot have rapid deployment of complicated functions and strategy without risk. Here, it is useful to create guardrails to reduce the consequences of risk. These will look different for each business and team. 

Sidebar: managing your data intelligently

Get to value quicker by empowering a strong person with an overarching mandate that includes oversight of the following elements of the data lifecycle: 

Collection: make sure you get what you need by doing it yourself in close cooperation with your product and development department. This includes: 

Data pipelines.

Augmentation: transform the data you collect, and bake in context and intelligence to create actionable data, by doing the following: 


Activation: actionable data needs to be prepared and provisioned to work for you. Make sure you include:


Execution: once provisioned and primed, marketers and the mar-tech they operate take over to reach out to your customers. Elements to consider here:

Manual vs automation.
Optimisation and testing.

Feedback: consistent feedback of engagement statistics feeds your models and triangulates your customers’ positions in the various conversion funnels – which serve as triggers for your journeys. These include: 

Event-level activity.

That is a mouthful. Let us stop here for now. In the next instalment, we will discuss how to develop a clear route to value for your data strategy. This will include how to offset your liability and risk with the potential benefits, and what these benefits look like.