Last week in Nairobi, Dentsu Kenya unveiled a new consumer intelligence “stack” — combining its Consumer Connection System (CCS) data hub, the Merkury audience platform, and the COPO (Consumer In Your Pocket) AI tool — that promises to transform how brands market in Kenya and across Africa. The CCS is a unified African consumer dataset blending digital campaign performance metrics, a proprietary survey panel, telco/partner data and clients’ own data into a single source of truth. 

On top of that sits Merkury (for Media and Enterprise), a person-based identity and activation platform that can push audience segments into major channels (Meta, Google, TikTok, etc.) via cookieless and clean-room integrations. Finally COPO is a mobile‑first AI interface that uses generative AI to simulate “synthetic personas” of real African consumers in real time, letting planners test ideas and get instant insights.

Together, Dentsu says the CCS, Merkury, and COPO stack enables precision targeting, cross-channel planning and rapid learning loops at an unprecedented scale for the region. In pilot campaigns, Kenyan banks, insurers, FMCG brands and even betting and alcohol advertisers saw dramatic uplifts: for example, a banking A/B test showed 4.4× higher click‑throughs and 3.6× more engagement (2.50% vs 0.57% CTR) with Merkury audiences, capturing 43.3% of all leads from a smaller audience pool as seen on slides presented during the launch event last week. 

This is a potential game-changer for marketing in Kenya and the rest of Africa. To date, many Kenyan and African campaigns have been hampered by fragmented data (separate teams for media, CX, creative), basic interest-based targeting and slow reporting. Dentsu’s stack attempts to bridge those gaps with an Africa‑built data foundation and AI-driven insights, moving real people (not proxies) to the center of campaign planning. 

In a nutshell, Dentsu touting it as “the future of marketing in Africa” — but it also raises new questions (about data quality, representativeness and privacy, for instance) and practical challenges (in integration and cost). Therefore, I will aim to break down what Dentsu announced, how it works, why it matters for marketing in Kenya and the rest of Africa, real use cases and results, as well as risks and next steps for marketers. 

What Dentsu Announced

Dentsu Kenya formally introduced a new consumer intelligence stack comprising three layers:

• Consumer Connection System (CCS): First unveiled regionally in 2018, CCS is Dentsu’s proprietary African consumer dataset. It “combines four aspects of data” — actual campaign performance metrics, a proprietary consumer panel (Dentsu said it surveyed 3,600 Kenyans across urban and peri-urban areas), third‑party partner data (e.g. telco and other sources) and clients’ own first‑party data — to build a holistic view of audiences. CCS covers demographics, psychographics, life stages, brand preferences and even Out‑of‑Home media maps (with urban billboards geocoded). During last week’s launch event, Joel Rao (Co-CEO, Dentsu Kenya) emphasized that CCS “creates a single, privacy-compliant source of truth” on African consumers — a rich dataset “built since 2007” on authentic local behavior rather than global proxies.
• Merkury (Audience Platform): Dentsu’s Merkury is a person-based identity and audience platform (with Media and Enterprise product suites) built on that data. Using the CCS base plus integrations with client databases, Merkury creates persistent individual IDs that unify online and offline records. It then builds fine-grained audience segments (“clusters”) tailored to Africa’s markets (Kenya, etc.) — for example the banking list includes “Everyday Banking” (emerging earners, mass market, digital-first users), “Borrowing & Credit” (homeowners, vehicle buyers), “Wealth & Investment” (affluent professionals, HNWIs, investors) and so on as seen on slides from the launch event). Dentsu says Merkury can address 20+ million unique Kenyan profiles across platforms like Meta, Google, TikTok and Snapchat, tying clients’ first-party data into that system. Crucially, Merkury integrates with enterprise tools (Adobe, AWS, Snowflake, Salesforce, etc.) to activate in a privacy-safe way: for example, a Merkury Clean Room allows brands to match customer lists with platform audiences without exposing raw PII.
• COPO (Consumer in Your Pocket): The newest layer is COPO, Africa’s first mobile-first AI platform built on verified consumer truth. COPO is a chatbot/interface that lets marketers query the CCS/Merkury data in plain language, at any time. Behind the scenes it uses generative AI (with Retrieval-Augmented Generation or RAG) to simulate real Kenyan consumers and focus-groups in real time. Marketers can ask COPO questions like “What motivates our digital-first banking segment?” or “Which creative resonates with retirees?” and get instant, data-grounded responses — allowing real-time synthetic persona research and creative “stress-testing” long before a campaign launch.

In launching these, Dentsu highlighted partnerships to bolster the stack: integration with cloud and CDP infrastructure (AWS and Snowflake for clean rooms), and with major ad platforms (Meta, Google, TikTok, even Amazon’s ad network). Dentsu also noted COPO’s African rollout (already live in South Africa, Nigeria and Ghana, with Kenya coming online) and RAG technology so that every answer “is grounded in verified African data”. In short, Dentsu is pitching a fully integrated data-driven ecosystem (built in-house for Africa) for campaign planning, activation and analytics.

How the Stack Works

At its core, Dentsu’s new stack is data + AI + activation:

• Data Sources & Identity: CCS ingests multiple data streams. In the presentation, Joel Rao described CCS as merging digital performance data (what happened in past campaigns), a nationally representative consumer panel, partner data (e.g. telecom or survey sources) and clients’ own CRM or transaction data. All this is stitched into person-level profiles via Merkury’s identity graph. Merkury ID “unifies your known and unknown customers…to a single, persistent person ID”, combining offline and online identifiers. This means if Jane M. sees a billboard then later clicks a Facebook ad, both events can link to the same profile (if deterministically matched). CCS also enriches profiles with attributes like demographics, interests, media habits and purchase behavior, and even maps out outdoor ad locations (longitude/latitude). The result is a cross-channel “map” of Kenyan consumers, rather than siloed platform data. Importantly, all matching is done in a privacy-compliant way — Dentsu notes CCS is “privacy-compliant” and that COPO’s backend never shares one client’s data with another.
• Precision Audience Building: As the data is stitched, Merkury can rapidly build precision segments. Teams at Dentsu gave examples: merging panel insights and third-party data allows, say, an “Urban Female Entertainment Lover” audience that spans radio listeners and Facebook users at once. Instead of crude interest targeting, Merkury audiences combine behavior and survey signals. Once segments are defined, they’re activated via APIs and clean-room connections: Dentsu partners with Meta/Instagram, Google Ads, TikTok Ads, and even Amazon DSP so that these proprietary audiences can be used in DSPs or social campaigns. So a marketer’s first-party list and CCS insights can create lookalike or custom audiences on those platforms, without exposing the raw data externally. (Dentsu’s presentation at the launch event noted integrations with Adobe, AWS, Snowflake, etc. as part of Merkury.)
• Real-time Insight with COPO: The COPO layer makes all this a living, interactive process. Instead of waiting for weekly reports or new surveys, COPO turns the CCS/Merkury data into a chatbot. This is an AI using RAG (Retrieval-Augmented Generation) — meaning every answer it gives is backed by real Dentsu data. For example, one could ask, “How do our segments in Nairobi consume media?” or “Which new product idea would appeal to the ‘busy family’ segment?” COPO would spawn a synthetic persona (“Wambui, 34, urban mother of 2…”), simulate a conversation or focus group, and report aggregated insights (motivations, emotional drivers, likely response). Dentsu’s presenters at the launch event stressed that COPO works on real-time synthetic persona simulations — essentially AI focus groups at scale. This lets marketers “test campaigns before spending a cent”. This a radically different way of doing things from my perspective! 
• Privacy & Security: During the launch event, Dentsu also went to great lengths to emphasize that COPO and Merkury are built with privacy in mind. Data in COPO is siloed per client and never shared across organizations; conversational queries query the proprietary database without exposing raw data. COPO’s architecture keeps all training data internal. Clean-room technology (e.g. Snowflake) means even when matching client lists to ad platforms, the raw PII never leaves the safe compute zone. Because COPO uses RAG on Dentsu’s own African data (not ChatGPT-style web scraped content), every insight is “grounded in verified African data”.

In practice, the stack allows cross-channel planning. Planners can map a consumer journey from TV/OOH to mobile to e-commerce, using the CCS data and OOH map, then deploy an omnichannel campaign via Merkury Media. Insights flow back continuously: COPO can take campaign results and update persona profiles (e.g. a persona’s sentiment shifts if an ad flops). Summing up the flow, one slide showed a loop: CCS data feeds Merkury Enterprise (ID+ segmentation) and Media/Activation (allocate budget across channels), whose results then feed back into CCS. The ultimate goal is that each campaign “learns” faster, requiring less guesswork and with automated data orchestration from start to finish.

Why It Matters in Kenya & Africa

To appreciate the impact, consider the status quo. Marketers in Kenya and the rest of Africa have long faced fragmented data and slow cycles. Audience insights often live in separate teams (media, creative, customer experience), and decision-makers rely on crude metrics like reach or impressions. Market research surveys happen infrequently (every few months), and digital campaigns still often target by broad interests (e.g. “soccer fans” or “mobile banking users”) with little local nuance. Reporting lags can mean missing consumer shifts, and measurement across offline-to-online channels is spotty. In this context, Dentsu’s stack could be transformative in several ways:

• Precision Targeting vs. Interest Hunches: Instead of guessing (or relying on global targeting buckets), brands can target real Kenyan segments identified by data. For example, the banking campaign case study showed that using Merkury audiences (based on behavior and panel data) achieved 4.4× higher CTR (2.50% vs 0.57%) and 3.6× more engagement than a campaign using generic interest-based lists. The Merkury-driven ads captured 43.3% of all leads even though the interest-based pool was much larger. In plain terms, a target-specific approach can reach people who are truly in-market (not just “might be interested”) — making every ad shilling count for more. Dentsu trials in banking and FMCG saw 57.2% more total interactions compared to regular campaigns, and programmatic CTRs stayed above 5% (versus the 1–2% typical benchmark). Put bluntly, Merkury is turning paid social into a precision awareness engine rather than a billboard. 
• Faster Learning Loops: Traditional campaigns are slow to adapt; weeks of data collection might be needed before deciding on creative or spend adjustments. COPO aims to collapse that cycle. Since it can simulate consumer feedback instantly, agencies can test multiple ad versions or value propositions with synthetic personas and see which ones “resonate.” This “test-before-launch” capability is rare as AI personas let marketers get faster, credible insights without waiting on a traditional research timeline. In Kenya and the rest of Africa, where markets evolve quickly and budgets are precious, reducing days or weeks off research can avoid costly ad mistakes. During the launch event presentation Denstu claimed COPO drove about a 15% uplift in campaign efficiency (through better targeting and creatives) and roughly 40% faster audience segmentation, which directly accelerates time-to-market.
• Cross‑channel and OOH Integration: Dentsu pointed out that Kenyan consumers see brands via TV, radio, OOH billboards, mobile, etc. Yet most analytics focus on just one or two channels. By mapping consumers across media (even literally plotting billboards on a map) and tying that to digital profiles, Dentsu’s stack encourages true multichannel planning. For example, a planner could use CCS to find which radio and OOH ads drove spikes in website traffic, then feed that into Merkury to refine a social campaign. This breaks silos. It also solves a key measurement gap: one slide noted that “out-of-home media and retail road mapping” is now part of CCS, meaning no touchpoint is out of reach (no pun intended).
• Cost Efficiency & Improved KPIs: Combining the above, the new stack can make each marketing shilling go farther. By eliminating wasted impressions and by quick-learning, early evidence suggests much better ROI. In banking, the targeted asset-finance ads not only got more clicks but higher-quality leads than interest-based ads did. In the insurance sector, Dentsu said Merkury audiences yielded better click-throughs and lower cost-per-action than before (though exact numbers weren’t disclosed, presenters emphasized “significant cost savings” with data-driven segments). FMCG campaigns also benefited: a case of a men’s grooming product on TikTok/Meta saw unusually high video completion rates and engagement at a low cost, by reaching precisely defined lifestyle segments. Even in conservative categories like alcohol, Dentsu showed that using data clusters (e.g. “Premium Lifestyle Seekers” vs “Responsible Moderation” drinkers) yielded higher completion rates on digital ads.

In summary, the advantages over the old way are clear: precision versus broad-brush targeting; speed versus slow insight cycles; and unified planning versus fragmented silos. Dentsu’s presenters at last week’s launh event emphasized that these have real numbers behind them — for instance, one slide explicitly showed 4.4× CTR and 3.6× engagement lifts in one test (with 2.50% vs 0.57% CTR). We have to take some of that as Dentsu’s claim that campaigns with Merkury audiences are achieving well-above-average performance.

That said, these claims still need real-world verification beyond the agency’s pilots. The key point is that such a stack — if all parts work as advertised — disrupts the status quo of how media is planned and measured in Kenya and across Africa. It parallels global shifts (post-cookie, identity focus, AI creative testing), but is tailored to African realities. In practice, early adopters in Kenya will likely get a jump on these trends, turning the “human truth” of the market into actionable campaigns faster than competitors.

Use Cases & Sector Impact

Dentsu has already sketched out how the stack applies across major verticals. Each sector has its own custom clusters and success stories. In short, Dentsu is applying Merkury and COPO across the usual ad‑tech hotspots: banking/finance, insurance, fast-moving consumer goods, gambling/sports betting, alcohol/beverages (and e‑commerce). Each has bespoke segments and preliminary results. The consistent theme is: data-driven segments beat generic interest targets, boosting efficiency and engagement. For example, the betting use case transformed paid social into a precision engine (3.3M impressions, thousands of clicks) simply by layering Dentsu’s data. These serve as practical case studies that digital and media planners will watch closely.

Risks, Limitations & Open Questions

Now, while the new stack is powerful, marketers in Kenya and Africa should be realistic about the following challenges:

• Data Quality and Representativeness: Merkury’s performance depends on its underlying data. Dentsu touts “200 million+ verified profiles” in Africa, but in Kenya specifically it showed only 9.7M IDs in one slide (about 85% of Kenyans on mobile, by Dentsu’s count). How representative are these profiles? The panel is 3,600 people, which is good but may not capture all niches. If urban elites dominate the panel, rural or informal segments might be under-modeled. Poor match rates (when linking first-party data to the database) could also limit scale. Marketers should ask Dentsu for details on panel demographics and ID resolution rates before relying solely on it.
• Scale vs. Reach Tradeoffs: Hyper-targeting inevitably shrinks the reachable audience. For example, in the bank case, Merkury’s high-precision segment captured 43% of leads but was a smaller pool than interest-based targeting. In a country like Kenya with 50M people but still uneven digital reach, marketers must balance scale and precision. Overly narrow segments could exhaust quickly or incur frequency issues. It’s an open question how granular you can get before incremental gains plateau.
• Privacy and Compliance: Although Dentsu insists on privacy safeguards, there are still concerns. Kenya has a Data Protection Act (2019) and various guidelines on digital marketing. Any AI-driven persona tool (COPO) must be carefully governed: for example, if one asks about sensitive attributes (religion, health), what data does COPO use? Dentsu says conversational data is not accessible to others, but clients should verify how long user queries are stored and how the RAG model is kept up-to-date. There’s also the question of third-party cookies: Merkury solves this with its own identity graph, but global regulations (or platform policies) could change how easily first-party data can be integrated. Agencies must ensure consumer opt-in and transparency when building these rich profiles.
• Vendor Lock-in and Cost: This stack is proprietary to Dentsu. Brands worried about dependence on a single partner will want clear ROI. Subscriptions and agency fees for such a sophisticated stack could be high. The slides wryly admitted “solutions always come at a cost.” Marketers should scrutinize the commercial terms: do they own the audiences built? How portable are segments if they switch agencies? What are the long-term commitments (e.g. multi-year data building)? And cost is not just money — adopting COPO/CCS requires training teams in AI methods.
• Competition and Alternatives: Dentsu claims uniqueness (“only network with African data”), but other players (local or global) are also advancing data solutions in Africa. Google, Meta and telecom companies have their own insights tools, and local startups offer CDPs or ad targeting. Marketers should compare Merkury/COPO to typical data management platforms (CDPs) or DSP audiences.
• Hype vs. Reality: Lastly, COPO’s generative AI is on-trend, but as always with AI: do the synthetic personas truly reflect the market? Early promises of AI in marketing sometimes overshoot. Ultimately, COPO’s insights should be validated against real consumer feedback to avoid “AI hallucinations.” Marketers must treat it as one input — albeit a fast one — not gospel truth.

Practical Guidance for Marketers in Kenya & Africa

If you’re a CMO, media buyer or marketer considering these tools, here are some steps:

1. Audit Your Data Strategy: What first-party data do you already have (CRM, app usage, website, sales)? Map this to see what gaps Merkury/CCS could fill (e.g. offline behaviors, panel insights). Know your current KPIs (CTR, CPA, lift).
2. Pilot with a Test Campaign: Don’t overhaul everything at once. Dentsu’s launch demos involved A/B tests, which is smart. Run a side-by-side test on a smaller budget: one wing using traditional targeting, the other using a Merkury audience/COPO-informed creative. Measure engagement, cost, and leads. Use the promised metrics (e.g. CTR lift, cost per lead) as a benchmark.
3. Governance & Compliance Checklist: Ensure all data is handled legally. Review Kenyan Data Protection guidelines. Check how COPO will handle any personal data in queries. Verify that Dentsu’s clean-room and RAG setup meet your corporate data security standards.
4. Integration Planning: Since Merkury plugs into platforms like AWS, Snowflake, Salesforce and advertising APIs, check your tech stack compatibility. If you already use e.g. AWS or Snowflake, find out how Merkury’s clean-room will fit. If you use Adobe Analytics or another CDP, ask how data flows into Merkury (there are connectors).
5. Define Success KPIs: Before deploying, decide what “success” looks like. It could be a percentage increase in CTR or engagement, cost per acquisition, or time saved in audience creation. Track these pre- and post- implementation. Also monitor qualitative outcomes — e.g. did creative teams find COPO insights actionable?
6. Incremental Rollout: Start with the most digitized channel (like social ads) where you have clear tracking. Later expand to TV/radio/OOH planning. Use COPO for quick surveying of target personas before investing big in TV creative.
7. Procurement Tips: Negotiate pilots or proof-of-concept periods. Ask Dentsu for case studies relevant to your sector. Since the stack is new, try to cap costs or have “pay for performance” clauses on efficiency gains.
8. Team Training: Train planners and analysts on the new tools. COPO’s value depends on good questions — train marketers in “AI prompt” techniques. Also align your media buying and creative teams so they speak the same language of data.

Strategic Implications

Beyond individual marketing tactics, this launch has broader signals:

• For Agencies: Dentsu is staking a claim to be the lead agency in pan-African data/AI marketing. Rival agencies may need to build or partner for similar stacks which Dentsu’s Joel Rao noted that they are open to doing so in Kenya as they have already done so in other instances. Agencies without such a system risk being outgunned in offering “full-funnel” analytics. It may accelerate M&A or partnerships (e.g. more acquisitions of African data startups or deals with cloud AI firms).
• For Publishers: If targeting becomes people-based (not cookie-based), local media owners need to embrace clean-room tech and audience segments to stay relevant. TV, radio and OOH sellers should consider how they share data to be part of CCS/Merkury. Out-of-home providers, for instance, might integrate their location data with Dentsu’s mapping.
• For Platforms (Meta, Google, TikTok): These players have been moving in this direction anyway (with their own APIs for first-party data). Merkury effectively wraps all platforms into one interface, which could threaten ad inventory fragmentation. Platforms may push back on data sharing or change targeting rules; we may see new partnership terms emerge.
• For Martech/Vendor Ecosystem: This raises the bar for other CDPs and identity providers. Vendors like Snowflake or Adobe see an opportunity — Dentsu’s use of Snowflake clean rooms, for instance, highlights that technology’s value in Africa. We might see more Snowflake/AWS projects in marketing here. Conversely, global CDP vendors might look to integrate African panel data (a potential partnership with Dentsu itself).
• For Regulators: Kenyan regulators and data authorities will probably watch how this system handles PII (Personal Identifiable Information) and consent. They may issue new guidelines on AI in marketing (especially if COPO becomes widely used). Pan-African regulators could consider frameworks for cross-border consumer AI tools, since COPO spans multiple markets.

Conclusion & Next Steps

Dentsu Kenya’s unveiling of CCS, Merkury and COPO is a bold step into data/AI-driven marketing for Kenya and the rest of the African continent. It crystallizes a vision where marketers no longer rely on blunt instruments (cookies, interest tags) but on a living, local data model and instant consumer dialogue. If it works as promised, it could leapfrog long-established markets in digital sophistication.

In the coming months, we’ll watch for more proof points. How will businesses and brands in Kenya pilot these tools? How will SMEs (for whom data access is usually limited) get to benefit? How will competitors respond with their own platforms? For now, we have a clear technology stack and a set of compelling case studies — and therefore potentially a new frontier of data-driven and AI-enabled marketing in Kenya and the rest of Africa.