Here is the uncomfortable pattern that emerged across Turkish retail in 2020: a company opens more channels, invests more in martech, and yet the customer experience becomes more fragmented. A shopper who bought in-store last week receives an abandoned cart email. The loyalty app sends a discount for a product already delivered. The call centre agent has no visibility into the previous WhatsApp exchange. This is not a technology problem — it is an architecture problem. And the growing conversation around CDP (Customer Data Platform) is a direct response to it. But there is a critical distinction that most of the vendor narrative obscures: deploying a CDP and having a unified customer profile architecture are not the same thing. Confusing the two is how organisations end up with an expensive data warehouse replacement that does not move the needle on actual customer experience.The thesis worth defending: a CDP without identity resolution capability and a downstream orchestration layer produces no meaningful omnichannel outcome. It produces segment lists. Segment lists are useful — but they are the output of 2015 martech thinking, not 2021 omnichannel thinking. The structural difference is this: in segment-based campaign logic, a marketer defines a cohort (‘customers who have not purchased in 30 days with a basket value above a threshold’), exports a list, and fires a campaign. The list is stale almost immediately. In profile-driven orchestration, the system reads the current state of the individual — which channel they are on, what they just did, what their predicted intent is — and triggers the next best action accordingly. The distinction is between the static nature of a segment and the dynamic nature of a profile. Getting from one to the other requires two things that no license agreement provides: clean identity mapping across systems, and an orchestration layer that can act on profile signals in near real time.A mid-sized home textiles retailer based in Izmir, with approximately 317 employees, illustrates the gap concretely. During the pandemic lockdowns of 2020, the company accelerated its e-commerce channel and began using WhatsApp Business for order updates and post-purchase communication. The loyalty programme, ERP inventory system, and e-commerce platform each held a separate customer record. When the marketing team ran a cart abandonment campaign, it reached customers who had already purchased in-store that same week — the e-commerce platform had no visibility into POS transactions. The campaign’s true impact could not be measured because ‘abandoned cart’ was defined differently in each system. This was not a CDP problem. It was an identity resolution problem that a CDP deployment would need to solve as a prerequisite, not as a by-product. The sequence matters enormously: many organisations acquire the CDP first and discover the identity fragmentation problem only during implementation, at which point the project scope expands significantly and timelines shift by six to eight months.The move from segment-based to individual-based orchestration carries a specific set of operational requirements that are frequently underestimated. First, there must be a mutually agreed definition of customer identity across marketing, sales, customer service, and digital product teams — which email address, phone number, loyalty code, and device identifier constitute the same person. This internal negotiation typically takes weeks and is considerably more difficult than the technical configuration that follows. Second, the orchestration layer needs to receive profile signals and trigger channel actions within a timeframe that is meaningful to the customer interaction — an abandoned cart signal that arrives 72 hours late has a different conversion profile than one acted on within four hours. Third, consent management must be embedded in the architecture before data unification begins. Under Turkish data protection law (KVKK), consent granted for a loyalty programme does not automatically extend to digital advertising retargeting. When this mapping is not completed before technical build, the unified profile immediately creates legal exposure. This is not a theoretical risk — it has forced scope revisions in multiple CDP projects in the Turkish market.The Turkish context adds a layer of financial complexity that the international CDP vendor narrative does not address. Licence costs for the leading international platforms are dollar or euro denominated. In the currency conditions of early 2021, this translates into a meaningful share of a mid-market marketing technology budget on an annual basis. Domestic alternatives in the Turkish market generally cover segment management and e-mail integration competently, but real-time identity resolution and cross-channel orchestration remain at an earlier maturity stage. This creates a genuine dilemma for organisations that need omnichannel capability but cannot absorb the foreign currency exposure of enterprise-tier international platforms. One pragmatic approach observed in the field: use a domestic CRM or marketing automation tool as the orchestration interface while building identity resolution capability separately, on a data infrastructure layer that the organisation actually controls. This is not a perfect architecture, but it is a more honest starting point than an enterprise CDP deployment that the team cannot yet operationalise.A word of caution on the measurement side. The omnichannel benefit case is often made with e-mail open rates or click-through rates on orchestrated campaigns. These are not the right metrics. The correct measure is cross-channel conversion contribution — specifically, how much revenue can be attributed to a customer journey that touched more than one channel before converting. Measuring this requires a multi-touch attribution model to be defined and implemented before the orchestration layer goes live. In the Ankara-area furniture retail environment, a team that ran individual-level orchestration saw engagement rates approximately 35 percentage points higher than their prior segment campaigns — but when they tried to quantify the revenue contribution across channels, they found their attribution model had not been designed for cross-channel journeys. The engagement metric looked compelling; the business case remained partially unproven. Attribution architecture is the part of CDP projects that gets deferred most consistently and costs the most in credibility when results are reported to leadership.Three actionable starting points for anyone evaluating CDP in the current environment. Map your existing customer identity fields first: how does your e-commerce platform, your POS system, your loyalty programme, and your CRM each identify the same customer? If there are more than two or three non-overlapping identifier fields, a dedicated identity resolution exercise is required before any CDP selection decision makes sense. Second, resolve your KVKK consent scope with legal counsel before the technical architecture is finalised — specifically, which data unification operations require which consent type, and how that maps to your current opt-in data. Third, write at least one concrete use case that describes how the CDP will trigger a channel decision, not just populate a segment list. Without that use case defined in operational terms, the project will drift toward becoming a data consolidation initiative rather than an orchestration one. The Izmir retailer mentioned at the outset is still building this capability. Their systems are in place. Their data is not yet clean enough to trust. That gap — between having the platform and having the prerequisite — is where most CDP projects currently sit.
This article was originally published in Turkish by Gökhan MERCANOĞLU on April 19, 2021. The English edition has been reviewed and edited by the author.