Mark Arnold has been working on customer relationships for 30 years. In his role at Havas CX Helia as Head of Data Consulting, he’s well qualified to assess the state of customer engagement and data strategy.

Helia’s heritage is in CRM and loyalty, with its previous guises including EHS Brann (Brann being Christian Brann, described by Helia as ‘The Godfather of Direct Marketing’ who before setting up his company in the ‘60s had previously run Reader’s Digest’s UK mail order business).

I ask Arnold what changes he has seen in his field over the past decade, from the focus on attribution in digital media, which some argue led to an unhealthy short-termism, to the rebalancing effect of new regulation and privacy frameworks.

“There’s a sense of over complication in the markets,” he says. “I’ve been through different iterations… it’s gone from direct marketing, through to CRM, through to customer engagement, through to CX.”

“The channels and technology that we can explore for [customer engagement] have grown… But the principles behind it are all pretty simple, and have never changed from the days of direct marketing. If you look at the language that we use, everyone talks about ‘right place, right time’. That’s not new, it’s the same language we used 30 years ago.”

There’s a sense of over complication in the markets.

“All we’re trying to do is have a meaningful dialogue with someone. Meaningful enough to encourage them to change the way they think or their behaviours.”

On the movement away from third-party tracking, he’s a big advocate, both from the point of view of the consumer and the business.

“I’m a fan of the fact that it places more power in the hands of consumers and I’m a fan of the fact that you, (the brand), need to get your experience right within your own channels to negate the need to be looking for those people elsewhere. Make your own experience so much better that they’re going to revisit your brand and your own channels.”

“What do you want the consumer to do?” – Back to the basics of CRM

The result of some marketers becoming, as Arnold puts it, “lost in the technology, in the data… excited about everything that we can do,” is that “they start building journeys and communications programs without a sense of purpose.”

He says he and his colleagues often talk to CRM teams who find it hard to articulate the purpose of individual emails. Marketers are too often looking for ‘engagement’, whereas Arnold asks them, “What do you want the consumer to do?”

“So if you take a welcome journey, it isn’t just to welcome people into the brand. That’s one purpose of the journey, but actually, you’re saying to them, ‘what value has that journey delivered to customers and the brand in the short term or long term?’

“Have you tried using a fallow group, a hold out, a control, and can you prove the incremental value that your welcome journey delivers to the business?

“…you may not see any shifts for a while, but you might further down the line, you know? You might be establishing that nurturing program and bringing them into the brand and getting them to think, early on, about specific behaviours you want them to take down the line. You need to be able to prove that.”

Email and eCRM - Econsultancy's Internet Statistics Database

“It’s all clicks” – Marketers relying on engagement metrics

Arnold’s team spans the Havas CX network, working hand in hand with strategists, planners, tech teams, and with insight teams, showing the degree to which data has become the fulcrum of the marketing department.

“Our role really in data strategy is about helping clients understand more about audiences and customers through insight and analytics. Looking for opportunity within the customer database, the data attributes needed, supporting on market research, and building out what we call the DNA of the customer.

“We’re also heavily concerned with measurement …We’re working more and more with clients in our advertising agency, Havas London, to help them demonstrate impact on brand metrics and how brand can influence demand metrics, lower funnel metrics, and vice versa.

“And the other thing my team do is take responsibility for using martech. Demonstrating how we can build out data strategies and… use that to create better experiences [by] activating that data and those audiences into the martech.

Lost in the technology… [marketers] start building journeys and communications programs without a sense of purpose.

When his team takes on a new piece of work for a brand, a maturity assessment is often the first job to be undertaken. “…we’ll have a look at performance, we’ll go through a discovery process… Just have a look at the lay of the land. Look at the benchmarks to see where they are today,” says Arnold.

He adds that, a lot of the time, when you ask for information to be shared, particularly around performance, “It’s all clicks. It’s all about engagement.”

And some of this also can be unreliable. For example, Arnold highlights that often brands are tracking email open rates with no real consideration for whether their customers are using iOS or other operating systems (this is pertinent because Apple users may have turned on Mail Privacy Protection, which hides their IP address and prevents senders from gathering this information).

On the other hand, for the data strategist looking to make the most of what they’ve got and prove the impact of a marketing campaign, Arnold sounds a positive (and pragmatic) note.

“People think that we can only [look at brand metrics] with pulse trackers and surveys, and brand trackers, [but] there are always other ways to measure things that you think you can’t see.

“[If you’re] trying to understand whether what you’re doing at a campaign level [has an] impact on brand, for instance, there’s a whole set of indicative metrics that you can look at, that you can use to infer whether you’re having an impact. If it’s about awareness… then you can be looking for the level of impressions and the amount of people that you’ve got following you. You can look at share of voice. Some of those are longer term behaviours as well.”


The integration of measurement and the ‘butterfly effect’

This role of measurement in finding the connections between activity and a disparate range of metrics is something Arnold returns to frequently, calling it the ‘butterfly effect’. He describes Helia’s use of a single dashboard approach, in an attempt to create a holistic view and measurement framework.

“On one side, you’ve got… awareness, consideration. …how people perceive the brand and all the brand’s associations… which again, influences everything that’s going on at the campaign level.

“…at the bottom you may have your program [or] campaign KPIs… you’ve got a set of customer KPIs and at the top, you’ve got your business KPIs. So, everything is laddering up, everything you do at the bottom… ladders up to what’s happening at a strategic level.

“And then you’ve got your data and insight KPIs. So, your ability to build models that are effective… Creating a quality score around your data.

“And now we’re adding another dimension, which is… sustainability, carbon impact scores… We need to look at them all on the one page on a dashboard. Quite often… you can see how something at a tactical level has started declining and [there’s] almost a dotted line… all the way up to what’s happening at the strategic level and can be having an impact on customer value, which in turn is having an impact on revenue and EBITDA. It lets you know where you can start turning the dials.”

I’m not seeing the integration of measurement.

This holistic view is something Arnold sees as lacking in many organisations’ approach to data and analytics. He describes “siloed teams” focusing on particular KPIs “within their team(s)”. “I’m not seeing the integration of measurement,” he adds.

“Everyone talks about [integration in] digital transformation in the guise of integration between agencies coming together to create these multi-touchpoint experiences, but they also require siloed teams to work together within an organisation and bring digital and CRM and social, and everything together.”

Data and Analytics - Econsultancy's Internet Statistics Database

AI tools – beware limitations, or lack of strategic context

At the Festival of Marketing way back in 2017 there was much talk of an AI summer. Big data, faster processors and open source algorithms were transforming martech, and AI was becoming a go-to selling point. Jim Sterne, author of ‘Artificial Intelligence for Marketing: Practical Applications’, said at the time that AI is the “most misused” term in the industry.

Six years on, this may sound quaint, in the middle of a news cycle that is peppered with pronouncements about generative AI and the potential threat of artificial general intelligence. But, it seems machine learning is still misunderstood, or the term misused, in the marketing community.

The FTC warned in February of this year, in a blog post titled, ‘Keep your AI claims in check’, that companies promoting AI in their products should ask, “Are you exaggerating what your AI product can do? Are you promising that your AI product does something better than a non-AI product? Are you aware of the risks? And does the product actually use AI at all?”

When I ask Arnold about AI in customer engagement, he echoes some of this sentiment.

“My jury’s out on AI to a degree. I think there’s a lot of hype about it. I think the term AI gets bandied around for anything, even when it could be quite simple. It’s being used as a bit of a sales term…

“[These tools might be] tracking engagement levels and building ready-made engagement scores and those sorts of things… And if you’re not working with an agency, you don’t have a budget, and you need a ready-made thing that’s going to help you across your team, that’s great, but actually they miss a whole lot of insight and analytics.

“We build propensity models, do what we call exploratory data analysis, just looking for value and opportunity within someone’s customer base. We look for engagement, we look to build RFVE (recency, frequency, value, engagement) models and things like that and we’re looking for headroom… and opportunity that those models don’t necessarily pick out.

“Those models, they’re not bringing all of the data together… They look at engagement but they can’t look at what’s happening to engagement at a transactional level. They don’t go deep enough.”

[If] you need a ready-made [tool] that’s going to help you across your team, that’s great, but actually they miss a whole lot of insight and analytics.

On generative AI, Arnold has concerns about the reliability of LLM-based tools for analysis, but says, “they are great at being able to review a corpus of information that you may have within a walled garden. For instance, if you’ve got a whole set of documents and insight and everything’s been captured. You might need an analyst or someone to walk through to pull out the meaningful information and story. [Generative AI] can read that for you and pull out the key information for you. So it’s great for those sorts of things. And it’ll be a really powerful use-case going forwards.”

He adds that within business intelligence, rather than requiring and asking someone within a team who may not be fully skilled to review dashboards and charts, which is prone to risk, generative AI may be a powerful ally.

But, Arnold adds, “There’s a strong risk that people… lose the strategic context” if decision engines are placed in charge of customer engagement.

Artificial Intelligence (AI) - Econsultancy's Internet Statistics Database

Pragmatic personalisation

Arnold’s realism extends to customer engagement and personalisation in the round. He describes the pitch commonly heard within martech or CX – “multichannel experiences, every touchpoint with consistent content” – but says, “it’s much more difficult than it reads on paper, and [that goes for] measurement as well”.

“You need to take small steps… in your roadmap and prove the value of each individual [step]. You’re adding to that until you get to [a] real understanding.

“[I] see so many business cases trying to get the investment internally to put into their data and everybody’s [asking for] 20% uplift in engagement. But how much of that do you really need? And how much [personalisation] is really demanded by different audiences?  …Where is the bar?”

Whether the topic is martech, AI tools, measurement and tracking, or the fundamentals of CRM, Arnold shows the value of asking clear-eyed questions. Data literacy and data strategy are crucial for marketing teams, particularly as they struggle with both long- and short-term challenges. A glance at the 2023 Marketing Week Careers and Salary survey shows that more than a third of marketers say their team is lacking data and analytics skills. More training, more external expertise, and more knowledge amongst leadership is clearly needed.

Further resources: