Free Website Audit: Discover what's holding your digital presence back
Short answer
A customer data platform should bring approved customer data together so a business can understand a person or account and take an appropriate action. Before buying or building one, define the decisions, source systems, identity rules, consent, access, activation needs, ownership and maintenance. A platform is not a substitute for clean processes or responsible data use.
Customer data platforms can sound like a single product category, but the business problem comes first. A company may need a shared customer view, more reliable audience segments, better service context or safer activation across systems.
Start with the decision you want to improve. Konzept’s software development service can help assess integrations, data flows and custom requirements when an off-the-shelf platform does not answer the full problem.
What should a customer data platform solve?
A customer data platform should make a defined customer or operating decision easier, more accurate or more timely. It may help teams understand a customer’s relationship across channels, coordinate a service response or create an audience for an approved campaign.
Write the decision before listing features. Examples include identifying a returning customer, suppressing a message after a purchase, helping support see relevant account context or giving a sales team a reliable lifecycle stage.
If the business cannot say what action follows the data, a platform may become an expensive reporting layer. Start with one valuable workflow and examine what information it truly requires.
Which data sources belong in the platform?
Include sources that are relevant to the decision, governed by your company and technically reliable enough to use. A source may include a website, ecommerce platform, CRM, support tool, app, offline record or consent system.
Map each source before integration:
| Data question | What to document |
|---|---|
| What is the source? | System owner, purpose and access method |
| What does the field mean? | Definition, format and allowed values |
| How fresh is it? | Event, batch or manual update timing |
| Who owns the truth? | System of record for the field |
| What can the team do with it? | Consent, contract and policy limits |
| What happens when it is wrong? | Correction, deletion and escalation path |
Do not connect every system because an integration exists. Unnecessary data creates more mapping, storage, access and maintenance work.
How should identity resolution work?
Identity resolution should define how records from different systems are matched and what happens when the match is uncertain. Decide which identifiers are reliable, which combinations are allowed and when the system must keep records separate.
Keep a distinction between a person, an account, a device and an interaction. A business buyer may act for several accounts. A household may share an address. A cookie or device signal may not prove a person’s identity.
Document merges, splits, corrections and deletion requests. A wrong connection can affect a message, service decision or report. The safest model is not always the one that creates the largest unified profile; it is the one that supports the decision without overstating what the data proves.
What governance does a CDP need?
A CDP needs clear ownership, purpose, access rules, retention, consent handling, security and a process for corrections. Decide who can create segments, export data, change mappings and approve activation.
Involve legal, security, marketing, sales, product and support where their work depends on the data. A platform does not remove the need for a lawful and understandable reason to collect and use information.
If you serve customers in Bosnia and Herzegovina and the EU, document which requirements and contracts apply to the relevant processing. Keep the user-facing explanation consistent with what the systems actually do. Avoid collecting a field just because a future campaign might want it.
Should you buy or build a customer data platform?
Buy when an existing platform covers the required sources, identity rules, governance and actions with acceptable control and maintenance. Build or extend when the workflow, data model, hosting requirement or integration pattern is genuinely specific and the company can own the system after launch.
Compare more than subscription features:
| Decision factor | Buying | Building or extending |
|---|---|---|
| Start point | Existing product and connectors | Custom design and integration work |
| Fit | Constrained by product model | Tailored to defined requirements |
| Ownership | Shared with vendor | Internal technical responsibility |
| Change process | Vendor roadmap plus configuration | Your roadmap and maintenance |
| Risk | Dependency, migration and vendor terms | Delivery, security and operational capacity |
Do not compare a vendor demo with a vague custom project. Write the required flows, data owners, controls and support model first. A custom software assessment can help scope the actual integration work.
How do you activate data responsibly?
Activate data only for an approved purpose, audience and channel. Define what happens when consent changes, a record is deleted, a segment expires or the source data conflicts.
Keep a trace from source to segment to action. A marketer should be able to explain why a person or account entered an audience and how to remove it. Test suppression, correction and failure paths before a campaign or customer workflow depends on them.
Use the ecommerce industry context when the platform supports product, order and service decisions, but do not assume every ecommerce event belongs in one profile. Keep the data model tied to the customer experience and business responsibility.
What should happen after implementation?
Plan for monitoring, documentation, access review, schema changes, vendor changes and user training after launch. Assign an owner for the platform and an owner for each critical source. Review whether the original decision improved before adding more integrations.
When the current systems, data definitions and governance are ready for a technical roadmap, book a software discovery call with the decision you want to improve. A smaller trusted data flow is often a better starting point than a platform that tries to represent everything.
FAQ
Does every business need a customer data platform?
No. A business may solve its immediate need with a well-governed CRM, analytics setup, integration or service workflow. A CDP becomes useful when several systems need to support a shared customer decision and the company can own the data, access and maintenance responsibilities. Define the problem before adopting the category.
What is the biggest risk when implementing a CDP?
The biggest risk is treating connected data as automatically correct, permitted or useful. Poor identity rules, unclear ownership, excessive collection and weak access controls can create harmful decisions at scale. Start with a defined workflow, document sources and consent, test edge cases and assign people who can maintain the system.
How long does a customer data platform project take?
There is no honest fixed timeline. It depends on the number and condition of source systems, identity rules, governance, integrations, required actions and internal availability. A focused first workflow can be scoped separately from later expansion. Define the initial decision and acceptance checks before estimating the broader platform.