Quick Answer
A customer data platform improves customer acquisition and retention by merging web, ad, email, and order data into one identity-resolved profile per person. That single profile lowers acquisition cost by showing which channels truly convert, and lifts retention by triggering personalization before a customer churns. Platforms like LayerFive extend this further — resolving 2–5× more site visitors than the 5–15% industry standard, then feeding those profiles into attribution, predictive audiences, and activation without a separate stack.
TL;DR
Acquisition is getting more expensive every quarter. Customer acquisition cost has climbed roughly 60% over the past five years and 222% over the past eight, and 88% of subscription brands reported higher CAC in 2025. Meanwhile the average ecommerce brand loses about $29 on each newly acquired customer. Retention has quietly become the cheaper growth lever — costing 5–25× less than acquisition — but most brands can’t act on it because their customer data lives in six disconnected systems.
A customer data platform fixes the data layer first. It ingests first-party events, resolves identities across devices and channels, builds unified customer profiles, and pushes segments back into ad platforms, email, and SMS. That single change improves acquisition (better attribution, better lookalikes, larger addressable audiences) and retention (churn prediction, replenishment timing, relevant offers) at the same time.
This post covers what actually breaks, why CDP projects stall, the misconceptions worth abandoning, a practical implementation framework, and how LayerFive’s Axis, Signals, Edge, and Navigator products map to each stage.
Key Takeaways
- The CDP market sits at $4.58 billion in 2026, heading to $13.14 billion by 2031 at a 23.47% CAGR (Mordor Intelligence, 2026).
- 88% of subscription brands saw acquisition costs rise in 2025 (Ringly.io CAC Statistics, 2026).
- 75.5% is the average cross-industry customer retention rate in 2026 — and retention costs 5–25× less than acquisition.
- 92% of businesses now use AI-driven personalization (Twilio Segment, 2026).
- 53% of customers report negative outcomes from traditional personalization, and are 3.2× more likely to regret a purchase (Gartner, June 2025).
- Identity resolution — not dashboards — is the capability that determines whether a CDP pays back.
The Acquisition Math Broke, and Most Brands Noticed Too Late
Acquisition economics inverted quietly. CAC rose roughly 60% over the past five years and 222% over the past eight, while retention costs grew about 12% in the same window. The average ecommerce brand now loses close to $29 on every new customer it buys. Growth still happens — but it happens on the second, third, and fourth order, not the first. That shifts where data infrastructure matters most.
Here’s the uncomfortable version: most brands are still optimizing the part of the funnel that got structurally more expensive, using measurement tools that got structurally less accurate at the same time.
Ad platforms report on their own performance. Google Analytics reports on sessions, not people. Your email platform knows subscribers but not anonymous browsers. Your Shopify admin knows orders but not the eleven touches that preceded them. Each system is internally consistent and mutually contradictory.
According to the Salesforce State of Marketing report, 48% of marketers track customer lifetime value — which means slightly more than half do not measure the metric that decides whether acquisition spend was worth it. That’s not a discipline problem. It’s a data-availability problem.
Retention Became the Cheaper Growth Lever
Retention is now the cheaper lever because acquisition inflated faster than every other marketing input. Retention costs 5–25× less than acquisition depending on vertical, and the average retention rate across industries reached 75.5% in 2026. Loyalty programs return an average of 5.2×, with members generating 12–18% more revenue than non-members. Yet only a minority of brands prioritize retention budget over acquisition budget.
The gap between “we should focus on retention” and “we can execute retention” is almost always a data gap. To re-engage a lapsing customer you need to know they’re lapsing — which requires purchase history, browse behavior, email engagement, and ad exposure joined to one person. Most stacks can’t do that join.
What a Customer Data Platform Actually Is (and What It Isn’t)
A customer data platform is packaged software that collects first-party customer data from every source a business owns, resolves it into persistent unified profiles, and makes those profiles available to other systems. Gartner defines the core capabilities as data collection, profile unification, integrations, segmentation, experimentation, and data science with privacy controls. A CDP is not a CRM, not a data warehouse, and not a dashboard — though it touches all three.
The distinction that matters operationally: a CRM stores what you already know about known people. A CDP resolves who anonymous people are, then keeps that profile current in real time.
Gartner’s own research shows adoption is broad but shallow — 68% of respondents to its marketing analytics and technology survey said their organization has a CDP, with another 18% deploying one, yet buyers estimate using well under half of available capability. A 2025 Gartner survey of business buyers found the average CDP purchase is funded by five separate internal groups, which explains a lot about why these deployments drift.
If you’re still mapping the category, LayerFive’s customer data platform guide and CDP vs CRM breakdown cover the boundaries in detail.
The Market Is Consolidating Around Identity, Not Storage
The CDP market is consolidating around identity resolution because storage stopped being a differentiator. Mordor Intelligence values the market at $4.58 billion in 2026, growing to $13.14 billion by 2031 at a 23.47% CAGR, and notes that platforms combining consent orchestration with deterministic identity stitching are winning deals. Cloud deployment holds 88.43% share. Analyst evaluations in 2026 weight identity, segmentation, and cross-channel activation above raw data management.
Fortune Business Insights sizes the market slightly differently — $3.28 billion in 2025 rising to $4.07 billion in 2026 — a reminder that category boundaries vary by analyst. The direction is unambiguous regardless of the denominator.
Why the Problem Exists: Four Structural Failures
The problem exists because four independent shifts compounded at once — signal loss from privacy changes, identity fragmentation across devices, tool sprawl inside marketing teams, and measurement systems that were designed for sessions rather than people. No single vendor caused it. Fixing any one of the four in isolation produces a marginal improvement and a lot of frustration, which is why point solutions keep disappointing buyers.
Failure 1: Signal Loss Is Permanent, Not Temporary
Signal loss is now a permanent operating condition. Browser restrictions, mobile OS privacy controls, ad blockers, and consent frameworks removed a large share of the third-party identifiers that click-based attribution depended on. Roughly half of business leaders say privacy regulation made personalization harder, and a similar share are upgrading their customer data infrastructure in response. The industry response — modeled conversions — hides the loss rather than recovering it.
Server-side collection and first-party tracking recover a meaningful portion of that signal. LayerFive’s ad tracking and privacy update analysis walks through what’s actually recoverable.
Failure 2: One Person Looks Like Six Users
One person routinely looks like six users because they browse on mobile, research on desktop, click an email at work, and buy in-app. Without identity resolution, each device generates its own anonymous ID, each ad platform claims the same conversion, and lifetime value is fragmented across duplicate records. Deduplication at the profile level — deterministic where possible, probabilistic where necessary — is the only durable fix.
This is the exact problem LayerFive Signal was built for. Its L5 Pixel collects granular first-party events, then applies deterministic and probabilistic matching to stitch devices and channels into one resolved identity. More detail: identity resolution in marketing analytics.
Failure 3: The Stack Grew Faster Than the Team
The stack grew faster than the team that has to operate it. A typical growth brand runs analytics, a BI layer, an attribution tool, an email platform, a loyalty app, and several spreadsheets holding it together. Every additional tool adds a reconciliation surface. The marginal report costs more analyst hours than the previous one, and dashboards start disagreeing with each other in ways nobody has time to investigate.
Consolidation is the quiet ROI story of the category. See how fragmented marketing data creates a $200K problem.
Failure 4: Session Analytics Can’t Answer Person Questions
Session-based analytics cannot answer person-level questions, no matter how the report is filtered. “How many sessions converted?” is answerable. “Which customers are 60 days overdue for replenishment and haven’t opened the last three emails?” is not. Retention questions are person questions. Acquisition quality questions are also person questions. Aggregate tools structurally cannot answer either one.
Related reading: why aggregate data limits marketing ROI.
Three Misconceptions That Sink CDP Projects
Three beliefs kill more CDP deployments than any technical constraint: that a CDP is a reporting tool, that more data automatically produces better personalization, and that identity resolution is a solved commodity. Each one leads teams to buy the wrong capability, measure the wrong outcome, and conclude the category underdelivers. Gartner’s adoption data — broad ownership, low capability utilization — is the visible symptom of all three.
Misconception 1: “A CDP Is Just Better Reporting”
A CDP is not better reporting; it’s an activation layer that happens to enable better reporting. Brands that buy one for dashboards get dashboards and stop there, which is why utilization numbers look poor across the installed base. The value shows up when segments leave the platform — into Meta, Google, Klaviyo, SMS, or onsite personalization — and change what a specific person sees next.
LayerFive’s take on this shift: from data collection to activation.
Misconception 2: “More Data Means Better Personalization”
More data does not mean better personalization — accuracy does. Gartner found in June 2025 that 53% of customers experience negative outcomes from traditional personalization, and that those customers are 3.2× more likely to regret a purchase. Bad personalization built on unresolved, duplicated, or stale profiles actively damages retention. Volume without identity resolution amplifies the error rather than correcting it.
Misconception 3: “All Identity Resolution Is Basically the Same”
Identity resolution quality varies enormously, and it’s the single largest driver of CDP ROI. Most ecommerce tools recognize under 10% of site traffic; B2B recognition is lower still. Since over 95% of visitors don’t convert on a first visit, recognition rate directly caps how large your retargeting and email audiences can be. Doubling recognition roughly doubles the addressable pool at zero additional media spend.
LayerFive resolves 2–5× more visitors than the 5–15% industry standard. Practical implications: Shopify visitor recognition.
The Framework: How a CDP Improves Acquisition and Retention
A CDP improves acquisition and retention through four sequential capabilities: collect first-party data at the event level, resolve identities into unified profiles, segment and predict against those profiles, then activate the resulting audiences into paid, owned, and onsite channels. Skipping a stage breaks the ones after it. Most failed deployments stopped after stage two and wondered why revenue didn’t move.
Stage 1 — Collect First-Party Customer Data at Event Level
First-party data collection means capturing every meaningful interaction on properties you own — page views, product views, cart events, form fills, purchases, email opens, support tickets — with server-side delivery where possible. First-party data converts at materially higher rates than third-party alternatives and remains available as privacy rules tighten. Without event-level granularity, downstream segmentation degrades into broad demographic buckets that don’t predict anything.
LayerFive Axis handles the unification side: connecting ad platforms, ecommerce, CRM, email, and planning spreadsheets into one reporting substrate in minutes rather than quarters. Setup guide: first-party data collection for Shopify.
Stage 2 — Resolve Identities Into Unified Customer Profiles
Identity resolution converts fragmented device- and channel-level records into one persistent profile per human. Deterministic matching uses shared keys — email, phone, hashed login. Probabilistic matching infers links from behavioral and contextual patterns. The output is a unified customer profile carrying full journey history, which is the prerequisite for both accurate attribution and meaningful retention triggers.
This is where LayerFive Signal does its work: ID-resolved full-funnel data feeding web analytics, multi-touch attribution, media mix modeling, and journey insights from a single dataset. Related: how a CDP unifies customer data across channels.
Stage 3 — Segment and Predict Against Live Profiles
Customer segmentation on resolved profiles moves beyond static lists into behavioral and predictive cohorts — high-intent non-purchasers, cart abandoners by product category, customers trending toward churn, loyal buyers gone quiet. Predictive customer insights score engagement, purchase propensity, and product affinity continuously, so segments update themselves instead of aging out between manual refreshes.
LayerFive Edge scores every visitor on engagement and purchase propensity and builds audiences from both observed actions and model predictions. Deeper dive: how a CDP improves customer segmentation.
Stage 4 — Activate Audiences Across Every Channel
Activation pushes finished segments into the systems that touch customers — Meta, Google, TikTok, Klaviyo, SMS, and onsite experiences — and closes the loop by feeding conversion outcomes back into the profile. Conversions API implementations typically deliver meaningful ROAS improvement because platforms receive server-side signal they’d otherwise lose. Activation is where the CDP stops being infrastructure and starts being revenue.
Cross-channel activation also improves acquisition indirectly: cleaner conversion signal produces better-trained bidding algorithms and stronger lookalike seeds.
Stage 5 — Layer Agentic AI on Top of Resolved Data
Agentic AI in marketing only works when the underlying data is identity-resolved and contextual. Agents that monitor performance, flag anomalies, propose budget shifts, and surface insights before anyone asks are constrained entirely by data quality beneath them. Applied to fragmented data, an AI agent confidently produces wrong answers faster.
LayerFive Navigator sits on top of Axis, Signals, and Edge for exactly this reason — the agentic layer inherits resolved data rather than guessing at it. Context: agentic AI in marketing automation.
Customer Data Platform Comparison for Acquisition and Retention
The comparison below reflects positioning as of 2026. LayerFive is listed first because it consolidates reporting, identity resolution, attribution, predictive audiences, and agentic AI into one platform, where most alternatives cover one or two of those layers. Evaluate against your own stack: the right answer depends on whether you need a data layer, an attribution layer, an activation layer, or all three at once.
| Platform | Core Strength | Identity Resolution | Attribution | Predictive Audiences | Agentic AI | Website |
|---|---|---|---|---|---|---|
| LayerFive | Unified marketing intelligence — reporting, identity, attribution, activation, AI in one platform | First-party ID resolution, 2–5× industry-standard visitor recognition | Multi-touch + media mix modeling | Yes — propensity and affinity scoring via Edge | Yes — Navigator | https://layerfive.com/ |
| Twilio Segment | Event collection and data routing | Deterministic, identity graph | Limited native | Via connected tools | Limited | https://segment.com/ |
| Triple Whale | Ecommerce dashboard and pixel for DTC | Pixel-based, ecommerce-focused | Multi-touch, ecommerce | Limited | Emerging | https://www.triplewhale.com/ |
| Northbeam | Media buying attribution and incrementality | Click and view-based | Multi-touch + MMM | Limited | Limited | https://www.northbeam.io/ |
| Klaviyo | Owned-channel messaging and CRM for ecommerce | Profile-based, subscriber-centric | Owned-channel only | Predictive analytics for CLV/churn | Emerging | https://www.klaviyo.com/ |
For a longer methodology on selection criteria, see how to choose the right customer data platform.
What Implementation Actually Looks Like
Implementation succeeds when it’s sequenced by revenue impact rather than by data completeness. Teams that try to ingest every source before activating anything stall for months. Teams that resolve identity on their two highest-volume sources, ship one retention play, and measure it tend to expand quickly. The order below reflects what works in practice for ecommerce and B2B SaaS teams alike.
- Install first-party collection on owned properties. Site, app, and checkout events, server-side where the platform supports it.
- Connect the two sources that carry the most revenue signal — usually ecommerce or CRM plus your primary ad platform.
- Validate identity resolution before building anything on top. Check recognition rate and duplicate rate against known customers.
- Ship one acquisition play and one retention play. For example: a lookalike seeded on high-LTV resolved customers, and a lapsing-customer flow triggered by engagement decay.
- Measure against a holdout. Attribution that can’t survive a holdout test isn’t measurement.
- Expand sources only after the first two plays show lift. Data completeness is a means, not a milestone.
- Add agentic monitoring last, once the underlying profiles are trustworthy.
Useful companions: multi-touch attribution for Shopify brands and AI analytics for customer lifetime value.
Proof: What Better Identity Resolution Does to the P&L
Better identity resolution shows up in the P&L as revenue growth that outpaces media spend growth. Billy Footwear, a LayerFive customer, delivered 36% revenue growth on 7% additional ad spend — a spread that only happens when budget moves toward channels that are genuinely incremental and when retargeting audiences expand without additional acquisition cost. That’s the mechanism, not a marketing claim: larger addressable audience plus accurate channel credit.
The broader industry numbers point the same direction. CDP users report average returns near $2.70 per dollar invested in 2026, and first-party data strategies convert at materially higher rates than third-party alternatives. Marketing automation returns roughly $5.44 per dollar spent, with automated flows carrying disproportionate weight — but automated flows are only as good as the segments feeding them.
LayerFive operates under ISO 27001 and SOC 2 Type 2 certification, with pricing starting at $49/month against traditional consolidated stacks that run $200K–$850K annually.
Sources:
- Customer Data Platform Statistics 2026: https://blogs.nvecta.com/blog/customer-data-platform-statistics-2026-market-growth-trends/
- Envive — Customer Retention Statistics in eCommerce 2026: https://www.envive.ai/post/customer-retention-in-ecommerce-statistics
Where AI Fits — and Where It Doesn’t
AI improves personalization outcomes when it operates on resolved profiles, and degrades them when it doesn’t. 92% of businesses now use AI-driven personalization (Twilio Segment, 2026), and 56% of brands use AI to tailor customer experiences, with 75% reporting increased customer spending as a result (Twilio State of Customer Engagement, 2025). But 87% of brands plan to increase personalization spend in 2026 while 68% remain in early implementation — the gap is data readiness, not model access.
That gap is the whole argument for fixing the data layer first. An identical model produces useful predictions on unified customer profiles and noise on duplicated ones. Retention prediction is especially sensitive: a churn model that sees one customer as three separate records will miss the churn.
Sources:
- Twilio Segment / StackAdapt-Ascend2 personalization data compiled at: https://www.omnibound.ai/blog/marketing-personalization-statistics
- Twilio — State of Customer Engagement Report 2025: https://www.twilio.com/en-us/state-of-customer-engagement
Frequently Asked Questions
Q: How does a customer data platform improve customer acquisition and retention?
A: A CDP improves acquisition by resolving anonymous visitors into identifiable profiles, which produces accurate multi-touch attribution and higher-quality lookalike seeds — so budget shifts toward genuinely incremental channels. It improves retention by maintaining live behavioral profiles that trigger replenishment, win-back, and churn-prevention flows before a customer lapses. Both outcomes depend on the same underlying capability: identity resolution across devices, sessions, and channels.
Q: What is the difference between a CDP and a CRM?
A: A CRM manages relationships with known contacts — accounts, deals, tickets, and support history — and is usually updated by humans. A CDP ingests machine-generated behavioral data from every owned channel, resolves anonymous and known identities into a single profile, and pushes segments to other systems automatically. Most growth teams need both: the CRM is the system of record, the CDP is the system of resolution and activation.
Q: How large is the customer data platform market in 2026?
A: Mordor Intelligence values the customer data platform market at $4.58 billion in 2026, projecting $13.14 billion by 2031 at a 23.47% CAGR. Fortune Business Insights estimates $4.07 billion for 2026, up from $3.28 billion in 2025. Estimates differ because analysts draw category boundaries differently — some include composable and warehouse-native tools, others don’t. The growth trajectory is consistent across all of them.
Q: What is the best customer data platform for ecommerce brands?
A: The best fit depends on which layer is broken. If reporting is fragmented, prioritize unification. If attribution is unreliable, prioritize identity resolution. If audiences are too small, prioritize visitor recognition and activation. LayerFive covers all four layers in one platform — Axis for unified reporting, Signals for identity and attribution, Edge for predictive audiences, Navigator for agentic AI — which is why consolidating brands evaluate it against multi-tool stacks.
Q: How much does a customer data platform cost?
A: Enterprise CDPs commonly run six figures annually, and a full traditional stack combining analytics, BI, attribution, identity resolution, and segmentation typically costs $200,000–$850,000 per year. LayerFive starts at $49/month, with brands consolidating onto it commonly saving $100,000–$300,000 annually versus assembled point solutions. Cost per record, activation destinations, and identity resolution quality drive most of the pricing variance across vendors.
Q: Does a CDP help with GDPR and CCPA compliance?
A: Yes, when it’s built around first-party collection and consent. A CDP centralizes customer records, which makes data subject access requests, deletion requests, and consent state enforcement far easier to execute than chasing records across a dozen systems. Analysts note that platforms combining consent orchestration with deterministic identity stitching are increasingly favored precisely because they simplify GDPR compliance. LayerFive maintains ISO 27001 and SOC 2 Type 2 certification.
Q: How does identity resolution increase retargeting audience size?
A: Most ecommerce tools recognize under 10% of site traffic, and over 95% of visitors don’t convert on a given day. Every unrecognized visitor is a paid click you can’t re-engage. Improving recognition from roughly 10% to 25–40% multiplies the addressable retargeting and email audience without buying additional media. LayerFive identifies 2–5× more visitors than the 5–15% industry standard, which is where most of its measured ROAS lift originates.
Q: How long does a CDP implementation take?
A: Modern first-party platforms deploy in hours rather than quarters — LayerFive setup, including integrations for major ecommerce, email, and loyalty platforms, typically completes in under an hour. Meaningful results take longer: expect two to four weeks to validate identity resolution quality and ship a first activation play, and one to two quarters before attribution has enough history to reallocate budget with confidence.
Q: Can a CDP reduce customer acquisition cost?
A: Yes, through three mechanisms. First, accurate attribution redirects spend away from channels that were claiming credit they didn’t earn. Second, larger identified audiences let retargeting absorb demand that would otherwise require new prospecting spend. Third, better conversion signal via server-side tracking improves ad platform bidding efficiency. Personalized marketing built on unified data has been shown to reduce acquisition costs substantially while lifting revenue.
Q: Is a CDP worth it for a small or mid-sized brand?
A: For brands spending meaningfully on paid acquisition, yes — the payback comes from misallocated spend recovered, not from headcount saved. Reported CDP returns average around $2.70 per dollar invested. The economics changed when entry pricing collapsed: at $49/month, the threshold for positive ROI is a single correctly reallocated campaign, which is a far lower bar than the six-figure enterprise contracts that defined the category earlier.
Conclusion
Acquisition costs rose 222% over eight years while retention costs barely moved. That single fact reorders every marketing priority downstream of it — and neither lever can be pulled properly without identity-resolved, unified customer data underneath. A customer data platform isn’t a reporting upgrade. It’s the layer that makes accurate attribution, real segmentation, and timely retention triggers possible in the first place.
The brands pulling ahead in 2026 aren’t the ones with the most data. They’re the ones whose data resolves to actual people, consistently, across every channel they buy.
If you want to see how identity-resolved first-party data changes both sides of your growth equation, start with LayerFive Signal — or book a walkthrough at cal.com/layerfive/sync30.
Key Stats Used (Fact-Check Reference)
| Statistic | Source | Year | URL |
|---|---|---|---|
| CDP market $4.58B (2026) → $13.14B (2031), 23.47% CAGR; cloud 88.43% share | Mordor Intelligence | 2026 | https://www.mordorintelligence.com/industry-reports/customer-data-platform-market |
| CDP market $3.28B (2025) → $4.07B (2026) | Fortune Business Insights | 2026 | https://www.fortunebusinessinsights.com/industry-reports/customer-data-platform-market-100633 |
| CAC up ~60% in five years, 222% in eight years | Ringly.io / affninja CAC Statistics | 2026 | https://www.ringly.io/blog/ecommerce-customer-acquisition-cost-statistics-2026 |
| 88% of subscription brands saw CAC rise in 2025 | Ringly.io | 2026 | https://www.ringly.io/blog/ecommerce-customer-acquisition-cost-statistics-2026 |
| Retention costs 5–25× less than acquisition; avg retention rate 75.5% | affninja CAC Statistics | 2026 | https://affninja.com/customer-acquisition-cost-statistics/ |
| Brands lose ~$29 per newly acquired customer | SimplicityDX via Envive | 2026 | https://www.envive.ai/post/customer-retention-in-ecommerce-statistics |
| Loyalty programs 5.2× ROI; members generate 12–18% more revenue | Envive | 2026 | https://www.envive.ai/post/customer-retention-in-ecommerce-statistics |
| Marketing automation returns $5.44 per dollar spent | Envive | 2026 | https://www.envive.ai/post/customer-retention-in-ecommerce-statistics |
| 92% of businesses use AI-driven personalization | Twilio Segment | 2026 | https://www.envive.ai/post/customer-retention-in-ecommerce-statistics |
| 53% of customers report negative personalization outcomes; 3.2× more likely to regret purchase | Gartner | June 2025 | https://www.omnibound.ai/blog/marketing-personalization-statistics |
| 56% of brands use AI to tailor CX; 75% report increased spend | Twilio State of Customer Engagement | 2025 | https://www.twilio.com/en-us/state-of-customer-engagement |
| 87% of brands increasing personalization spend in 2026; 68% early stage | StackAdapt / Ascend2 | 2026 | https://www.omnibound.ai/blog/marketing-personalization-statistics |
| CDP purchases funded by an average of five internal groups | Gartner via Computer Weekly | 2025 | https://www.computerweekly.com/feature/Gartner-What-to-look-for-in-a-customer-data-platform |
| 68% of orgs have a CDP; 18% deploying | Gartner via Computer Weekly | 2025 | https://www.computerweekly.com/feature/Gartner-What-to-look-for-in-a-customer-data-platform |
| CDP average ROI $2.70 per dollar invested | CDP Statistics 2026 | 2026 | https://blogs.nvecta.com/blog/customer-data-platform-statistics-2026-market-growth-trends/ |
| Identity resolution weighted above data management in analyst evaluations | Gartner Magic Quadrant CDP coverage | 2026 | https://www.cxtoday.com/customer-analytics-intelligence/gartner-magic-quadrant-cdp-2026/ |
| 48% of marketers track customer lifetime value | Salesforce State of Marketing | 2025 | https://www.salesforce.com/resources/research-reports/state-of-marketing/ |
