Balancing The 3 S’s (Scale, Signal And Safety) In A New World

Paul Cimino

Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today’s column is written by Paul Cimino, president at Cimino Collaborative Consulting.

Safety and privacy have become driving forces in advertising. New laws, ad blocking and OS/browser changes have turned the industry upside down.

I sensed major changes coming in 2013, when I predicted the demise of cookies and see even bigger transformations on the horizon. To navigate these changes, brands must balance their desire for scale (reach) and signal (accuracy) with the growing requirements of safety (privacy).

Unfortunately these three S’s don’t play well together. They oppose one another strategically and sit in separate technologies. But advertisers must get the three S’s to work together across multichannel brand spend, and doing so requires massive changes to their advertising and data stacks.Currently ad tech maximizes scale and signal without accounting for privacy. For example, you can have scale (via cookie sync) and signal (slamming data together) but must sacrifice safety through cookies and daisy-chained opt-ins. You theoretically could have “safe” campaigns, with clear, transparent opt-in and signal, but the resulting scale would be close to zero. Nevertheless, while you can’t compete with the three S data mountains Google, Amazon and Facebook, you can still balance the three S’s for your brand and drive profit with an audience-centric strategy.

Why is it important?

Equilibrium among these three opposing forces is paramount because we are in the midst of the biggest change in advertising and media since the introduction of cable TV – that is, the switch from media-centric to audience-centric marketing. Media-centric means traditional media buying using CPM tracking, look-back attribution and macro connections between audience and media types. This model has begun to yield to audience-centric advertising and omnichannel journey mapping that allows for the merger of upper funnel (future customer, CPM) advertising and lower funnel marketing (current prospect, CPA).

Once this transition is complete, marketers stand to double overall ROAS by getting rid of that 50% #WannamakerWaste we’ve been talking about for 20 years.

Who is doing this successfully?

In short, nobody, because the tools are new and the know-how to glue all the data together and deploy an audience centric strategy is still two to three years away. But various brands are making progress, such as McCormick in CPG, Home Depot in retail and Bank of America in finance.

When should I do it?

Yesterday – this is a data/analytics war that has already begun. Brands need to hoard data, surround themselves with a moat of first-party tech and machine learn like there is no tomorrow.

What follows are guidelines for bringing the three S’s into harmony and alignment. By deploying this first-party data framework, advertisers can distill their unique audience and then use that unique data to drive profit – in effect, to corner their audience market.3S model chart 1

Solving scale

Most buy-side ad tech is built on islands. Scale is built by DSPs, signal is built by DMPs and consent is built by early-stage consent platforms. Although these systems are linked and synced, this is done through lose APIs and batch data matching processes which are prone to delays and poor accuracy.

In digital, brand reach is completely dependent on the API daisy-chaining of anonymous/pseudonymous IDs in the background pipes of ad tech. The demise of cookies and mobile IDs is a game changer. The cookie apocalypse effectively ends the near infinite spam-scale ad tech has enjoyed for 15 years. Once the dust settles advertisers will lose 60-80% of their ability to address media, target and measure campaigns.

To change this, brands and agencies need to either build their own programmatic bidder (on platforms such as or or they need to build a segmentation strategy in a machine learning environment (outside the DSP) and then integrate that ML method into a big platform, for example, The Trade Desk.

Solving signal

Marketers should use their first-party audience to define attributes that make up the best customers. But rather than dumping these attributes into a third-party look-alike modeler in a third-party DMP, they must keep this data in a first-party data lake, merge it with logs from their DSP, DMP and websites (all data) and then develop and refine a first-party enterprise machine learning application.

From here marketers can begin a process of refinement and optimization, adding more data and refining the number, weighting and formula between attributes. This supplemental data can come from a national file (Experian), panel data (Nielsen, Kantar, Fluent) or a multitouch attribution platform.

A “data moat” should also be constructed using a first-party ID/tag system. This identity graph and tag system should be under the domain/URL/cname and completely controlled by the enterprise. In this way, pubs and brands can avoid the need to trust their data to ad tech companies, a necessity given their history of abuse and the current regulatory risk.

That said, there is no magic first-party identity system that automatically links to the open web at scale. You still need partners. Your choices are:

  • Google, Apple, Amazon and Facebook clean rooms are data technologies that allow your data to interact with these partners’ data (match, analyze, merge) without exchanging data with partners.
  • Commercial consortiums. IAB Project Rearc and other consortiums seem to embrace a strategy of first-party ID “HEM cookies” anchored by ad tech luminaries such as LiveRamp, ID5, Tapad and so on. HEM cookies are the hashed email packaged with a consent string.
  • New entrants. There are too many to count but this group includes Snowflake, the data tech company that could upend the whole HEM cookie thing with the mother of all clean rooms and server to server linkage in your/their data cloud.

Solving safety

Similar to scale and signal, safety (privacy) is a first-party plus third-party framework. Safety has two main business cases: regulatory and personalization. Even though these are very different cases, the same platform must handle them.

First-party consent management is more about the ID systems and data structures and reporting systems than it is about a consent management application. That application should be sourced with a third-party provider, but should share all event/log level data with your data lake through your ID system. Leading third-party CMPs include, and, but note that most CMPs are sell-side focused and need augmentation.

The business case: Remember this isn’t only about an opt in or opt out, its about the five C’s of safety – clarity, choice, consent, control, children:

  • Clarity means your privacy notice has to be clear in terms of what you’re asking and what users get.
  • Choice means a simple opt-in and opt-out process.
  • Consent means your users understand what’s happening to their data on an ongoing basis.
  • Control means your users have a simple way to control their data.
  • Children means your Consent Management Platform needs to account for and manage tracking, targeting and messaging to kids as regulations dictate by state and by country.


To prepare for the future state of identity on the web, brands should aim for a mixture of first-party tech and best of breed partners around a central data powered, analytically-driven expert system that is internally owned and operated.

This enterprise marketing application set will become the ultimate decision maker for all future marketing deployment, merging the hunt for future customers and current prospects.

The CDP core components are first-party applications and data with the brands’ own unique business rules. These first-party components (identity graph, data lake, consent systems) are topped by an AI that is driven by many algorithms that act for the common good of the brand and its customers.

The challenge is constructing all these components economically and getting all the parts to work together around a central AI driven premise of increased profitability from lower cost (less waste) and higher return on ad spend.

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!