---
title: "Single customer view: business case, architecture, pitfalls"
description: "How to build a single customer view in Australia: the business case, 4 architectures, identity resolution, privacy rules and a 7-point data scorecard."
url: "https://greatminds.consulting/resources/blog/single-customer-view/"
language: "en-AU"
---

[Customer data & CX](https://greatminds.consulting/resources/blog/category/data-and-ai/customer-data-cx/) [Data integration](https://greatminds.consulting/resources/blog/category/data-and-ai/data-integration/)

# Single customer view: business case, architecture and pitfalls

A single customer view (SCV) is one trusted record for each customer: their identities, consents, transactions and service history, joined across every system and readable by every team. For most Australian mid-market organisations, the way to get there isn't a big-bang platform purchase. It's three decisions made in writing first: which system is the hub, how records are matched, and how consent travels with the data.

18 November 2026 · 14 min read

This guide covers the business case, the four architectures that work, how identity resolution actually behaves in Salesforce, Adobe and Klaviyo, the Australian privacy and spam rules that shape the design, and the pitfalls that stall most programs. It also includes a seven-point data quality scorecard adapted from the one standard in Australia that defines an SCV in law.

## What is a single customer view, exactly?

An SCV is a unified profile per person (or per business account in B2B) built from records held in several systems. It has four parts:

- **Identity:** every identifier you hold for the person, such as email addresses, phone numbers, loyalty and account numbers, and device IDs, linked to one profile ID.
- **Attributes:** name, address, preferences and segments, with a rule for which source wins when systems disagree.
- **History:** orders, bookings, cases, email engagement and web behaviour, each traceable to its source.
- **Permissions:** marketing consent per channel, opt-outs, and any restrictions on use, stored as data rather than as a note in a campaign tool.

An SCV is an outcome, not a product category. A CRM, a customer data platform (CDP), a marketing platform or a data warehouse can each host it. Master data management (MDM) is the discipline of governing the golden record; an SCV is what customer-facing teams and systems consume from it.

Australia does have one legal definition. APRA's [Prudential Standard APS 910](https://www.apra.gov.au/system/files/2023-09/Prudential%20Standard%20APS%20910%20Financial%20Claims%20Scheme%20-%20clean.pdf) (current version in force from 1 October 2023) requires every locally incorporated bank, building society and credit union to "identify each unique account-holder, to the extent practicable" and to generate an SCV of their protected balances. It sets hard expectations:

- The SCV must be generated within 48 hours of the relevant end of day.
- SCV systems and data must stay in Australia unless APRA approves otherwise.
- The account-holder identifier is capped at 18 characters.
- The CEO must attest each year that the systems produce complete and accurate data.

If you aren't an ADI, APS 910 doesn't apply to you. It's still the most concrete statement of what a reliable SCV looks like, and we use its data controls as the basis for the scorecard later in this guide.

## What's the business case for a single customer view?

The case rests on four levers. Revenue and cost benefits make it attractive; risk reduction and AI readiness usually get it funded in 2026.

Lever What the SCV changes What to measure (baseline before you start)

Revenue Segments built on full purchase and service history instead of one channel's data; triggers fired by real behaviour Revenue per contact, repeat purchase rate, conversion by segment

Cost Fewer duplicate records, so fewer wasted sends and lower profile-based licence costs; service staff stop searching three systems Duplicate rate, contacts billed by each platform, average handle time

Risk One place to honour an opt-out, correct a record, answer an access request or delete data you no longer need Time to action an unsubscribe across all systems, time to answer an access request, records past their retention date

AI readiness Agents and models read one governed profile instead of scraping inconsistent copies Share of AI use cases blocked by data access or quality

The risk lever has sharpened. The OAIC received [1,205 data breach notifications in calendar 2025](https://www.oaic.gov.au/news/media-centre/data-breach-notifications-increase-to-all-time-high-in-2025,-new-ndb-stats-show), the highest since the scheme began in 2018 and 8% more than 2024, with 716 caused by malicious or criminal attacks. Every uncontrolled copy of customer data adds to that exposure. The maximum penalty for a serious or repeated interference with privacy is now the greater of A$50 million, three times the benefit obtained, or 30% of adjusted turnover, according to the [OAIC's guide to regulatory action](https://www.oaic.gov.au/about-the-OAIC/our-regulatory-approach/guide-to-privacy-regulatory-action/chapter-7-privacy-assessments).

Spam enforcement is the more frequent cost. In April 2026 Latitude Finance [paid a A$3.96 million penalty](https://www.acma.gov.au/articles/2026-04/latitude-finance-pays-396m-more-spam-breaches) after the ACMA found more than 2.7 million spam contraventions between March 2024 and April 2025. Many of its texts told recipients to reply STOP when that reply couldn't work. When opt-outs live in one tool and sends go out from another, this is the failure an SCV prevents.

Proof that the revenue lever is real is easier to find in specific projects than in averages. When Great Minds migrated Meriton Suites from Salesforce Marketing Cloud to Klaviyo, the team moved more than 1 million contacts with their consent data, engagement history and segmentation attributes intact. It also connected the ecommerce and booking platform for real-time behavioural triggers. The [Meriton Suites case study](https://greatminds.consulting/resources/case-studies/meriton-suites/) reports a 4x increase in campaign execution speed and segmentation capability. We cover the migration itself in [Migrating from Salesforce Marketing Cloud to Klaviyo](https://greatminds.consulting/resources/blog/salesforce-marketing-cloud-to-klaviyo-migration/).

## How do Australian customers and regulators shape the design?

Australians want an organisation to know them, but not to over-reach. The OAIC's [2026 Australian Community Attitudes to Privacy Survey](https://www.oaic.gov.au/engage-with-us/research-and-training-resources/research/australian-community-attitudes-to-privacy-survey/australian-community-attitudes-to-privacy-survey2026/overview) (1,504 adults, fieldwork 16–30 March 2026) found:

- 93% say protecting their personal information is important to them.
- 78% feel they have very little or no control over how it's collected and used.
- 96% say selling or trading personal information isn't fair and reasonable (87% in 2023).
- 93% support a legal right to ask for their personal information to be deleted.

An SCV can be the tool that gives customers that control, or the database that makes their fears worse. Five rules turn the survey findings into design requirements.

- **Quality is a legal duty.** [APP 10](https://www.oaic.gov.au/privacy/australian-privacy-principles/australian-privacy-principles-guidelines/chapter-10-app-10-quality-of-personal-information) requires reasonable steps to keep personal information accurate, up to date and complete, and to check it is relevant before you use or disclose it. The OAIC's examples include auditing and correcting records, recording when information was collected, and giving people an easy way to update their details. A well-run SCV is how you show those steps.
- **Opt-outs must reach every system.** Under [APP 7](https://www.oaic.gov.au/privacy/australian-privacy-principles/australian-privacy-principles-guidelines/chapter-7-app-7-direct-marketing), you must stop direct marketing within a reasonable period after a request, generally no more than 30 days and shorter for digital channels. For email and SMS, the Spam Act applies instead, and the [ACMA's guidance](https://www.acma.gov.au/avoid-sending-spam) is to honour an unsubscribe within 5 working days. It also says to keep a record of who gave express consent, when and how. Consent therefore has to be a field on the profile, with its source and timestamp.
- **Delete what you no longer need.** [APP 11.2](https://www.oaic.gov.au/privacy/australian-privacy-principles/australian-privacy-principles-guidelines/chapter-11-app-11-security-of-personal-information) requires reasonable steps to destroy or de-identify personal information once it's no longer needed, including copies in archives and backups. Since 11 December 2024, APP 11.3 makes clear that reasonable steps include technical and organisational measures. An SCV that copies everything forever fails this test.
- **Know where the profile lives.** If a platform hosts profiles offshore, [APP 8](https://www.oaic.gov.au/privacy/australian-privacy-principles/australian-privacy-principles-guidelines/chapter-8-app-8-cross-border-disclosure-of-personal-information) on cross-border disclosure applies. Hosting options differ by vendor and plan. Adobe Experience Platform, which underpins Real-Time CDP, lists an [Australian region (AUS5, New South Wales, on Microsoft Azure)](https://experienceleague.adobe.com/en/docs/experience-platform/landing/multi-cloud); other platforms host Australian customers in the US or Europe. Ask every vendor in writing.
- **Misuse now has a civil remedy.** Since [10 June 2025](https://www.oaic.gov.au/privacy/your-privacy-rights/more-privacy-rights/statutory-tort-for-serious-invasions-of-privacy), individuals can sue for serious invasions of privacy, including misuse of information about them. Combining data in ways customers wouldn't expect is a bigger exposure than it was.

One more date matters for SCV programs that feed automated decisions. From 10 December 2026, privacy policies must explain when personal information is used in decisions made by computer programs that significantly affect individuals. If your SCV drives automated pricing, credit, eligibility or service decisions, read [Automated decisions and the Privacy Act](https://greatminds.consulting/resources/blog/privacy-act-automated-decisions/).

## Which architecture fits: CRM hub, CDP, or composable?

Four patterns cover almost every mid-market SCV. The right one depends on where your best customer data already sits and how fast you need to act on it, not on which vendor has the best demo.

Pattern Where the profile lives Fits when Watch out for

CRM as hub Salesforce Sales or Service Cloud (or Dynamics 365), with other systems syncing contacts into it Customer relationships are managed by people (B2B, wealth, professional services); fewer than about five source systems CRMs weren't built for high-volume behavioural events; storage and API limits rise with volume

Packaged CDP Salesforce Data 360 (formerly Data Cloud) or Adobe Real-Time CDP Many sources, high event volumes, real-time activation across several channels, and an existing Salesforce or Adobe stack Licence model and consumption costs; identity rules still need a business owner

Marketing platform profile A marketing platform's own profile store, such as Klaviyo's Direct-to-consumer and ecommerce brands where marketing is the main consumer of the SCV Service, finance and offline data often stay outside; hosting location may be offshore

Composable (warehouse-native) Your data warehouse (Snowflake, Databricks, BigQuery, Redshift), with tools activating from it A data team already runs a well-modelled warehouse and wants one copy of customer data Needs data engineering capacity; real-time use cases may still need a platform cache

The line between packaged and composable is blurring. Salesforce's [Zero Copy](https://www.salesforce.com/au/data/connectivity/zero-copy/) lets Data 360 query Snowflake, Databricks, Redshift and BigQuery without copying data in, and share its results back. Adobe's [Federated Audience Composition](https://experienceleague.adobe.com/en/docs/federated-audience-composition/using/overview) builds audiences from warehouse tables, such as Amazon Redshift and Azure Synapse, and imports only the audience into Experience Platform. Leaving data where it is helps with APP 11 as well as cost: fewer copies means fewer places to secure, correct and delete.

Whatever the pattern, the same seven layers have to exist somewhere:

- **Sources:** CRM, ecommerce, booking or point of sale, service desk, ERP and billing, marketing platform, web and app analytics.
- **Ingestion:** APIs, connectors or an integration platform moving data on a schedule or as events.
- **Identity:** match rules that link records to one profile, and reconciliation rules that pick the winning value.
- **Profile store:** the unified profile, with links back to every source record.
- **Governance:** consent by channel, purpose restrictions, access controls, retention rules and audit logs.
- **Activation:** segments and attributes pushed to email, SMS, ads, service consoles, websites and AI agents.
- **Measurement:** the baseline metrics from the business case table, tracked from day one.

Most failed SCVs had layers 1, 2 and 6 and skipped 3 and 5. Our [data integration](https://greatminds.consulting/capabilities/data-ai/data-integration/) practice usually starts there. For how the SCV fits with Atlassian, Adobe and ERP integration more broadly, see our pillar guide, [The connected enterprise](https://greatminds.consulting/resources/blog/connected-enterprise-systems-integration/).

## How does identity resolution actually work?

Identity resolution decides which records belong to the same person. Every platform does it with two kinds of rule. **Match rules** decide which records belong together, and **reconciliation rules** decide which value wins when the matched records disagree.

Match rules are either deterministic (an exact match on a strong identifier such as a customer number or a normalised email address) or fuzzy (similar names or addresses that allow for typos). Deterministic matching rarely merges two different people, but it misses some duplicates. Fuzzy matching finds more duplicates but risks over-merging, such as two family members at one address with the same initial. Over-merging is the worse failure, because it can send one person's details or offers to someone else.

Here's how three platforms Great Minds works with handle it:

- **Salesforce Data 360** uses an identity resolution ruleset. Salesforce's [Trailhead module](https://trailhead.salesforce.com/content/learn/projects/explore-data-cloud-core-functionality/unify-your-data) shows match rules combining fuzzy name with normalised email, phone or address. Reconciliation can choose the most frequent value, the most recent, or a value by source priority. Source records aren't changed; a unified link object traces every profile value back to where it came from.
- **Adobe Real-Time CDP** relies on Experience Platform's [Identity Service](https://experienceleague.adobe.com/en/docs/experience-platform/identity/home), which links two identities when their namespace and value match exactly, building an identity graph per customer. Matching is case-sensitive, so normalise emails before you send them in.
- **Klaviyo** merges profiles deterministically when one action links an email address, phone number or mobile device ID. Its [identifier hierarchy](https://help.klaviyo.com/hc/en-us/articles/12902308138011) runs Profile ID, then external ID, email, phone number and anonymous mobile app ID. Manual merges can't be undone.

Before you configure any of these, write a one-page match policy. It should name the identifiers you trust, in order; whether a shared phone number or address can merge two people; how you treat households and B2B contacts who move companies; and who approves changes. Then test it on a sample of real records, review the merges by hand, and only then switch it on for the full database.

## What does good SCV data look like? A seven-point scorecard

APS 910 lists seven controls an ADI must have over its SCV data. We've adapted them into a scorecard any organisation can run in a workshop. Score each line 0 (not in place), 1 (partly) or 2 (in place and tested).

# APS 910 control (Attachment A, para 6) The test for a marketing or service SCV Score (0–2)

1 Data sourced from any system is complete and accurate Each source has a named owner and a documented field mapping; load failures alert someone

2 All fields populated, or flagged as not available Missing values are explicit (unknown is not the same as no consent); required fields are enforced at capture

3 Calculated fields are accurate Lifetime value, last purchase date and segment flags reconcile to their inputs on a sample check

4 Name and address fields complete and valid Addresses validated against a reference file at capture; emails and mobiles checked for format and bounces

5 All records referenced to their source Every profile value can be traced to the system and record it came from

6 No duplicate records for a customer Duplicate rate is measured monthly; the match policy is written and approved

7 Balances reconciled to source systems Order, booking or revenue totals on profiles reconcile to the finance or commerce system each month

A score of 11–14 means you can build on what you have. A score of 6–10 means fix identity and ownership before buying anything new. A score of 5 or less means the SCV project is really a data quality project, and should be scoped and funded as one. Add one line of your own: can you action an unsubscribe in every sending system within 5 working days, and prove it?

## What are the most common single customer view pitfalls?

- **Choosing the platform first.** A CDP selection before the match policy and source inventory exist usually ends with the platform configured around whatever data was easiest to load.
- **No business owner for identity.** Match and reconciliation rules are business decisions. If only the implementer understands them, nobody can explain a bad merge to a complaining customer.
- **Over-merging to hit a number.** A falling duplicate count looks like progress. Review a sample of merges by hand before celebrating.
- **Consent left behind.** Migrating profiles without consent history, timestamps and source forces you to re-permission or to send at risk. The Meriton Suites migration kept consent data and engagement history with each contact for this reason.
- **Copying everything.** Loading every field from every system "in case" increases breach exposure and APP 11.2 deletion work. Bring in what a use case needs; federate the rest.
- **Real-time for everything.** Abandoned-cart triggers need seconds. Monthly segment refreshes don't. Real-time pipes cost more to build and run, so reserve them for use cases that pay for them.
- **Marketing-only scope.** If service, billing and booking data stay out, the SCV can't stop a promotional email reaching a customer with an open complaint.
- **No baseline.** Without before-and-after numbers for the metrics in the business case table, the program can't show its value at the first budget review.

## How long does it take, and how should you phase it?

Phase by use case, not by system. A useful first release can be scoped to one quarter if it is limited to two or three sources and one activation channel. Treat the timings below as a planning template; your source count and data quality score will move them.

Phase Indicative timing Scope Exit criteria

0. Discover Weeks 1–4 Source inventory, consent audit, scorecard, match policy draft, business case baseline, hosting and APP 8 review Architecture pattern chosen; first use case agreed and signed off

1. First use case Weeks 5–12 2–3 sources (for example CRM, ecommerce and marketing platform), identity rules, consent sync, one activation Use case live; duplicate rate and unsubscribe time measured against baseline

2. Extend Months 4–6 Service, booking or billing data; suppression rules across channels; service console view Customer-facing teams use the profile daily; scorecard at 11 or above

3. Activate AI Month 6 onward Grounding AI agents and models on the governed profile; automated-decision disclosures where needed Each AI use case has a data owner, access controls and a privacy assessment

## Where do AI agents fit in?

AI agents are only as good as the customer data they can read. A service agent that can't see a customer's last three orders, or a marketing agent that can't see an opt-out, does more harm than good. That's why the SCV has become the prerequisite for most customer-facing AI use cases rather than a separate project.

Two rules apply. Give agents the minimum profile fields each task needs, through the same access controls staff use. And log which data each agent used, so you can explain a decision if a customer or regulator asks. For use cases by department, see [Agentic AI for Australian mid-market: 40 use cases](https://greatminds.consulting/resources/blog/agentic-ai-use-cases-australia/).

## Frequently asked questions

### Do we need a CDP to have a single customer view?

No. A CRM or a well-modelled data warehouse can host an SCV, especially with five or fewer sources and limited real-time needs. A CDP earns its cost when you have many sources, high event volumes and several channels to activate in real time.

### Is a single customer view the same as master data management?

They overlap. MDM governs golden records across customer, product and supplier data, usually for operational systems. An SCV is the customer profile that marketing, sales, service and AI consume. Many mid-market organisations get an SCV without a separate MDM tool by using a CDP's or CRM's identity rules and a written match policy.

### Does our customer data have to stay in Australia?

For most organisations, no law requires it. APP 8 governs cross-border disclosure, and contracts, sector rules or customer expectations may require onshore hosting. ADIs are different: APS 910 requires SCV systems and data for Financial Claims Scheme purposes to stay in Australia unless APRA approves otherwise.

## Where to start

Start with the scorecard and a one-page match policy, not a platform shortlist. If those show you're ready, the platform choice usually becomes obvious from where your data already lives. Great Minds implements SCVs on [Salesforce Data 360](https://greatminds.consulting/partners/salesforce/data-cloud/), [Adobe Real-Time CDP](https://greatminds.consulting/capabilities/platforms/cdp/) and [Klaviyo](https://greatminds.consulting/partners/klaviyo/), and through our wider [customer data platform](https://greatminds.consulting/capabilities/platforms/cdp/) practice.

If you'd like a second opinion on your scorecard, match policy or architecture choice, [talk to our team](https://greatminds.consulting/contact/).

Talk to us

Want help applying this?

Our consultants can turn these ideas into a plan for your organisation.

[Contact us](https://greatminds.consulting/contact/)

Keep reading

## Related insights

[All insights →](https://greatminds.consulting/resources/blog/)

- [ Customer data & CX · 12 November 2026 Migrating from Salesforce Marketing Cloud to Klaviyo: lessons from 1M+ contacts How to move from Salesforce Marketing Cloud to Klaviyo: what the integration moves, Spam Act consent, active-profile costs, warming and cutover.](https://greatminds.consulting/resources/blog/salesforce-marketing-cloud-to-klaviyo-migration/)
- [ Customer data and CX · 9 November 2026 What is a CDP (customer data platform)? How it differs from a CRM A CDP unifies customer data from every system into one profile other tools can use.](https://greatminds.consulting/resources/blog/what-is-a-cdp/)
- [ Customer data and CX · 6 November 2026 Klaviyo and Shopify Plus for Australian retailers: data flows that hold up at peak season The Klaviyo and Shopify Plus data flows that fail at Black Friday: webhooks, consent, catalog sync, API limits, sending volume and Australian spam rules.](https://greatminds.consulting/resources/blog/klaviyo-shopify-plus-peak-season/)

## Let's work together.

Turn insight into action with a team that delivers.

[Work with Great Minds](https://greatminds.consulting/contact/)
