Customer Data Platform: Build Vs. Buy? Exclusive Expert Insights for 2020 and Beyond

Thanks to COVID19, we have seen a slew of virtual events and have even managed to attend more than we usually would have, what with all the commuting time saved. While many of these are informative, a recent conversation* between Stefano Opens a new window Fanfarillo Opens a new window , Partner and Director, Personalization & Digital Marketing at Boston Consulting Group (BCG) and Nick Rockwell Opens a new window , SVP Engineering at Fastly and Former CTO of The New York Times especially caught our attention, because it brings the focus to that most fundamental of questions around any technology: should we build it or buy it? In this case, the technology in question is customer data platforms, a Toolbox audience favourite.

^{*(Disclaimer: ActionIQ Opens a new window is the organizer of this seriesOpens a new window of virtual events featuring industry practitioner conversations. However, this is an independent editorial feature by the Toolbox team and is based on insights from the conversation.)}

When it comes to customer data platformsOpens a new window , it is more a question of ‘when and not if’, at least for marketers of a certain size that are talking to a significant base of prospects and customers. CDPs have proven to be one of the most effective technology platforms to empower data-driven marketers in an era of complex, multi-channel, personalization-led experiences.

The build-versus-buy topicOpens a new window around CDPs is of interest because this year, we are in an altered reality, and it would be interesting to check back in on what may have changed (or stayed the same), when it comes to making this decision for 2020 and beyond.

To set the context of the CDP build-buy discussion, BCG’s Fanfarillo’s description of the CDP does a good job of reminding us of its central capabilities. He defines it as ‘the technologies and processes’ to support marketing operations in four key areas:

1. Collate and integrate customer data across multiple sources
2. Create and maintain the 360-degree view of customer across all the interactions and transactions
3. Enable the development of customer insights
4. Support and orchestrate the activation of customer experiences across multiple channels and media

That is indeed a far cry from where CDPs began- essentially as a solution to store and manage data. In fact, CDPs still sometimes get bundled into conversations about data lakes (thankfully the confusion between CDPsOpens a new window and DMPsOpens a new window seems to have abated).

Today, established CDP vendors have moved far beyond data unification capabilities, into areas such as analytics, integration, and activation of the data to run orchestrated campaigns across multiple channels and devices. To paraphrase Stefano, a data lake is data at rest, while a CDP enables data to move.

What started out as an emerging technology championed by David Raab of the CDP InstituteOpens a new window was last year named by Gartner as one of three transformative marketing technologies for this decade (David Raab, Anand Thaker and I discussed the Gartner Digital Marketing Hype Cycle 2019 report at length with the author of the report, Mike McGuire from Gartner, in this episode of the Talking Stack PodcastOpens a new window ).

The number of vendors too has grown, along with robust funding; mergers and partnerships; and acquisitions of CDPsOpens a new window by ‘mainstream’ marketing, technology and advertising companies. A few examples include technology company Arm’s acquisition of Treasure DataOpens a new window , Mastercard’s acquisition of Session MOpens a new window and SaaS CMS Acquia’s acquisition of AgilOne.

We have also seen increasing product and capability specialization across vendors, including vertical specific solutions (fintech, healthcare etc.), size-specific solutions (CDPs for enterprises, SMEs etc.) and business-model specific solutions (B2C and B2B specialists). What this means is that if you decide to buy instead of build, the decision of picking a vendorOpens a new window that works for you is more complex than ever.

But if you are still at the build-versus-buy decision stage, let me preface this analysis with an important reminder from Nick Rockwell from Fastly, who labels the question a ‘false dichotomy’. He says, “In every build, there’s something that you’re buying – [most likely] foundational technology, such as the database or a workflow engine or a message vessel. And whenever you are buying, there’s also a build component, [like] integrations or some logic within the application platform”.

If you pare down the CDP to its roots and think of it as ‘scaled marketing automation’, then of course it is inevitable that everyone will have to do some version of it. A build-versus-buy decision for an advanced customer data managementOpens a new window technology such as CDP starts with asking yourself how much you are willing to invest to do that at scale and with sophistication, versus getting by with rudimentary or inflexible tools.

Necessary questions and avoidable traps

While there are a variety of dominant and supporting motivations that make you tend towards one approach over the other, start with the most strategic questions.

• How critical is data-driven and customer-centric marketing for your business?
• How sophisticated do your analytics and orchestration capabilities need to be?
• How critical is it to build, manage and leverage your first-party customer data capabilities at scale and speed?

Undoubtedly, building a CDP from ground-up is an undertaking fraught with risks. Not only are many marketers wary of getting into a long-term developmental effort with IT, they also are keenly aware of how expensive, inflexible and long-drawn the entire effort may be, often ending with obsolete functionality that no longer serves dynamic business needs. Unfortunately, aside from this awareness, Nick reminds us that marketers often tend to fall into three avoidable traps.

• The first is the tendency to feel that ‘our business is special and there are such unique characteristics or unique requirements that make it difficult to conceive of a third-party solution that could appropriately meet those requirements.
• The second is that because of the foundational data and analytics component of a CDP, it’s fairly easy to fall into a trap of thinking that you already have it. So, you think to yourself, ‘we already have the data, we already have some analytics platform, so maybe we almost already have our CDP.’
• The third is to approach the build as an IT-engineering led effort, disproportionately focused on solving for the technology layer, even before solving for core CDP capabilities such as connecting to different data sources, new activation technologies, or improving marketing ops effectiveness and efficiency.

The third point reminds us not to underestimate the criticality of other elements of the CDP, such as efficient UI for marketers to easily design and deploy campaigns, a marketer-friendly UX to drive adoption, as well as building a platform that can handle change, integration and orchestration with ease.

A marketer-friendly UI and UX

Says Rockwell, “Even with the best data and analytics, if you can’t quickly create, deploy and evaluate campaigns, you’re actually going to lose the value of the whole undertaking. The speed and lack of friction in that process is a critical component and a critical goal of the CDP.”

BCG’s Fanfarillo adds that he has seen homegrown, custom-built CDPs that can definitely integrate data, but don’t necessarily provide all the functionality and capabilities that a marketing user actually requires, in some cases even becoming a bottleneck to marketing operations. For example, adding a new attribute or connecting to a new activation engine needs dependence on an IT resource, which are often limited in terms of availability and support. “The whole solution”, he says “which was meant to be a key enabler for the marketing operations, turns out quite the opposite in some cases – solving for something but becoming a bottleneck in other cases.”

Change, integration and orchestration

Knowing how dynamic marketing channels are, and how complex today’s omni-channel customer journeys can be, the reality is that there are always new channels downstream that need to be integrated into the CDP. When thinking of building a CDP, a critical consideration is building a system that can integrate with new channels and tech stack components, as well as continue to deliver orchestrated campaigns with minimal disruption to the CX.

Consider the current scenario, where overnight, the entire channel of physical events and to a great extent, in-store shopping has been wiped out of marketers’ plans. How fast can your marketing systems adapt to this new reality and integrate new digital channels into marketing ops?

Rockwell, who is also the former CTO of The New York Times, suggests that it’s impossible to guess what channels may go or stay tomorrow. Instead, a good starting point is to establish your must-have channels for today and build for those first.

“You know that there will always be new channels and that they’ll always be evolving. So looking for depth and the right level of granularity in the channels that are most important to you as well as building an open ended architecture that will allow you to have channels in the future are both important.”

Adds Stefano, “The build is usually manageable, but the ‘maintain and evolve’ is a difficult thing to sustain.”

That said, there are still genuine considerations where a build decision makes perfect sense.

These include:

1. Level of customization and control: when a high degree of control is needed over the solution architecture. For example, highly regulated industries or those handling a depth of personal data. This also applies for companies who see their data as a strategic or proprietary asset that they do not want to risk sharing externally.

2. Organizational size: larger enterprises may indeed have the available resources and expertise to build themselves, and even be able to do so more cost effectively.

3. Data and tech comfort: enterprises in some sectors or industries have a greater comfort with data and using technology to manage that data optimally. For example, traditional financial service enterprises that have leveraged data analytics for years, versus some brick and mortar retailers, whose heritage may be in supply chain, not data.

Of course, even in the above scenarios, the ‘time-to-value’ is an important metric. Building may be less expensive if you have the resources in-house, but how much time will it take before the custom-built solution is able to deliver value?

Nick adds a reality check, “Builds are always hard. They will always take longer than you think. There’s always undiscovered scope you’ll find later. There’s always more work to do because a channel just changed how they operate and you have to redo the integration and so on. So, that’s really the filter: is there a good buy option? If there is, by definition that’ll get you to value faster.”

Unfortunately, buying is just as complex an exercise, with many nuances, and of course, potential vendors. A thumb rule is to approach the decision as a joint one between marketing and IT, and if possible, seek the advice of a neutral industry expert or systems integrator who knows and understands the technology, use-case scenarios and vendor ecosystem well.

Nick and Stefano both speak from experience when they say that with a buy decision, seek a vendor that has strengths in your weakest areas, and prioritise those with solid credentials in security and privacy to help go beyond regulations and actually enable a privacy-first approach to marketing.

Ultimately, whether you build or buy, a marketer-driven, use-case led, incremental development effort is key to deliver a successful CDP implementation.

If you would like to view this or other conversations in this ActionIQ Opens a new window series, register hereOpens a new window .