Marketers are unhappy with the level of integration between the different technology tools they use to do their jobs.
It’s an issue Which-50 has been interrogating with marketing executives, experts and analysts over the last three months, as we’ve tried to understand what’s driving a level of frustration with marketing technology solutions.
The research, which we conducted in partnership with Cheetah Digital, included a survey of 170 executives, along with 12 in-depth interviews with C-suite executives. The findings were released at a panel discussion in Melbourne in mid August.
- Download Our Strategy First report that describes the integration and other challenges CMOs are facing
- Sign up to Which-50’s Digital Marketing Newsletter
Marketers have assembled their technology stacks to deliver cohesive customer journeys at scale. But simplicity isn’t easy to achieve and those ‘seamless customer experiences’ are delivered on platforms built from disparate systems and data that have been stitched together.
Messaging from vendors offers the promise of a single customer view and the tools to deliver the right message at the right time to the right customer. And although they are making progress towards that goal, the holy grail is harder to achieve than the B2B sales pitches typically acknowledge.
CMOs Which-50 spoke to told us that even within the marketing clouds that sell the idea of a single unified solution, the truth is far more circumspect. Many marketing clouds are amalgamations of products acquired over several years, which do not always integrate well — even inside a vendor’s own cloud.
Robin Matlock, VMware CMO, told Which-50 earlier this year that while marketing cloud vendors like to talk in terms of universal stacks, the truth is that the marketing technology infrastructure that exists today is built from “pieces and parts that are not working together”.
As a technology CMO, Matlock has the benefit of understanding how hard it is to get everything working together seamlessly under the bonnet. She said the marketing technology industry still has a way to go to try to deliver on its promises.
“They put the narrative out that says the systems work together, but the code is separate, the architectures are different,” Matlock said. “They now need to integrate them. And I think the industry has a way to go to really deliver on the promise of the martech platform.”
That sentiment was also reflected in research conducted by Australian digital strategy agency, Ntegrity, released last week. Less than half of marketers speak completely positively about their martech stack and one in five report only negative things about it, with many complaining it is a convoluted mess or in a never-ending state of construction.
Anecdotal feedback included statements that reflect the challenges of a highly fragmented martech ecosystem, such as “Lots of bits cobbled together,” “It’s all over the place and is not highly integrated at this point in time,” “It is a bit of a mish-mash of tools that don’t fully integrate with each other,” and “Bolting things on here and there”.
We asked Scott Brinker, the editor of Chiefmartec.com — and now also VP of Platform Ecosystems at Hubspot — about the integration issue during his visit to Australia last week.
He said he saw similar research a year ago suggesting integration was still the number-one reason people gave for why martech wasn’t working for them.
“Part of me found this surprising. There are dozens of great iPaaS (integration platform as a service) solutions available out there, which are designed to make integration easier. What isn’t working?”
When he considered the issue, he said he realised part of the problem is that “integration” is too broad a term.
“Integration is a very high-level term. If you ask ‘do application A and application B integrate with each other?’, in today’s world, the answer is almost always yes. Now, exactly what that means, well, it turns out there’s a lot of different possibilities.”
Research from Gartner agreed that “integration” is a broad term used to describe myriad ways to sync data, content and business logic across systems.
“Marketing technology leaders strive for an integrated stack. However, no standard approach or mechanism exists to achieve integration excellence,” the analysts wrote.
Approaches to integration
As he broke the problem down, Brinker said he identified four different ways to look at integration (illustrated in the chart below).
“First up is just the exchange of data fields between two different systems. Now, even within such data-layer integrations, you can have a lot of variance. It might be a light integration, where any time you have some data to share, you just toss it over the wall to the other app. But what happens then? Was it accepted? Was there a conflict? Did it get merged with another record? Well, you don’t know.”
“Or you can have deeper data integration, involving a synchronisation mechanism, which does things like testing for conflicts and providing a process to resolve them, and which ensures changes get replicated in both directions.”
The growing popularity of workflow automation introduces a new layer, he said. “Okay, you’ve got all these different apps in your tech stack. It’s one thing to simply pass data between them, but increasingly, you often want to set up processes that trigger activities across those different apps.
“Do you have the ability to orchestrate workflow between application A and application B? That’s an entirely different kind of integration.”
The third level, he suggested, is integrating the user interface and user experience across apps.
“You might be exchanging data, you might be able to trigger events for workflow between two systems. But as a user, do you still have to jump back and forth between those apps to get things done within each of their interfaces, which are completely different from each other? That’s a disjointed experience with a pretty high cognitive tax on people’s day-to-day work.
“What you ideally want is visibility across both apps within the same window. So if I’m logged into App A, and I need to do something related with App B, that piece of App B’s functionality is contextually embedded within App A’s UI, right where I need it. I don’t need to go switching windows and having to reorient myself.”
And again, as with data, these UI and UX integrations themselves can be light or heavy, depending on how much functionality is embedded between them.
The ecosystem approach
The final layer of integration is what Brinker called the governance layer.
This is really where companies like HubSpot, that are in the business of managing ecosystems around their products, can bring value to customers by providing standards, certification, and oversight of the apps they publish in their marketplaces, he told Which-50.
“If you buy an app from one of our certified partners, you know that we’ve reviewed it, that the partner has agreed to our policies for operating as a ‘good’ app, and that we’re going to hold them to those standards. This is increasingly valuable around compliance requirements, such as GDPR.
“That’s actually very useful. In the absence of a well-structured ecosystem, in the Wild West, each vendor could shrug their shoulders and say ‘not my problem’. Good governance around an ecosystem makes the participants accountable and builds trust with customers.”
Playing nicely with the business systems
Integration challenges don’t just exist within the marketing department. Often, customer experience use cases will require data collaboration with IT, sales, product and other lines of business.
Simon O’Day, the Chairman and Co-Founder of The Lumery, said the challenges of integration are a natural next step to a modern marketer’s move into more and more data-driven work.
“For any direct marketer who worked from the ’80s-2000s in print and then email, the challenges of data and integrations are not new — they feel like they are getting easier and easier to deal with,” he told Which-50.
For example, O’Day noted the task of getting data from core systems, older versions of CRM and product data was usually considered ‘technical development’ that was expensive, specialised work.
“Today marketers are more than used to being data-driven, and so getting much more across the need for data to build their programs. And vendors have exploded in this area to help assist and make this more normalised.”
When asked about the disappointment Which-50 has observed in the market, O’Day said it is likely the marketer buying the technology hasn’t understood what data is actually needed to ‘fuel’ their program and where and how to get that data for use in the marketing technology.
“The easy answer is to blame salespeople but this is not a new challenge,” he said. “If someone has done a proper internal capability audit before embarking on using a martech stack in a modern way, then they will find the gaps in people, process, tech and the data needed — and factor them in.”
A bigger problem, O’Day said, is the ability for marketers to access data from the other systems in the business that paint a holistic picture of a customer’s profile.
“The largest problem that we see, day in and day out, is the inability for a marketer to access, structure and ingest data from the business core systems that house raw files, that enable a marketer to utilise it in a customer-friendly manner. Second to this is the lack of people who know how to address all of this.”
In practice, that could look like an email marketing program which is only fed names and email addresses, limiting the power of the platform to perform any greater level of personalisation.
“Suddenly so many of the use cases around customer experience and lifecycle are dead in the water and it’s back to batch and blast,” O’Day said.
“Not being able to get access and use of product data, core customer data or purchase behaviour data easily and in a timely manner crushes the program and spirit.
“This causes what we call the ‘shelfware effect’ — tonnes of money spent on tech now sitting still because nobody thought about the critical pieces.”
O’Day recommended that marketers do the work upfront — before speaking to vendors — to clearly identify the problems they are trying to solve and how those align with the business goals. They should have a clear view of the data, skills and budget required before seeking out a technical solution.
“From this most, if not all, integration and use case challenges will emerge and can either be solved or drive the outcome.”
Gartner research noted that vendors continue to make progress at packaging integrations into configurable connectors that non-technical marketers can use to connect data, content, business logic and experiences.
However there are few standards and numerous mechanisms marketers employ to accomplish connectivity with other systems, the analysts argued.
“The ability to effectively evaluate and make marketing technology decisions requires understanding different integration mechanisms — the tools, interfaces, and code that help tie together disparate systems and databases,” the analysts wrote.
For example, integrations could be achieved by APIs, prebuilt connectors or tags, snippets of code placed on a web site to collect data, set cookies or activate third-party content.
Integration could also be achieved by different kinds of platforms such as iPaaS, iSaaS or marketing technologies that natively facilitate integration such as a Customer Data Platform (CDP) or Data Management Platform (DMP).
Which approach marketers take will depend on relationships with technical teams, and their own use cases.
Gartner urged caution, and specific contract terms, when relying heavily on prebuilt connectors provided by vendors.
“A rich network of integrations may be a tempting draw, but don’t assume that vendors will continue to work together in perpetuity. If you go the prebuilt connector route, codify the expectations of continued integration support in your contracts and service-level agreements during procurement and renewal,” Gartner recommended.
O’Day said there will be fewer requirements for additional integrations in the core products of a stack from 2020 onwards, but it depends on when the actual components inside the cloud were introduced.
Nevertheless, the integration task is an ongoing process as new capabilities — responding to changing customer demands — come online.
“The big players in the market are always going to be both acquiring and building/improving their tech. This means this work is not going to go away fully but is a focus of improvement. This also means that there will always be some conflict from the tech sales team’s idealised visions and the actual use cases on the ground,” he said.