Jan 2016

adtech advertising

This is an internal essay from 2016 used to pivot my adtech startup Instinctive from a native ad network to a sell-side platform. Posted here for historical transparency. In hindsight, I missed the rise of header bidding that year with prebid.js and the major industry shift it caused which further changed the role of networks and SSPs.

Instinctive was founded in 2014 with a focus on creating a better advertising experience for advertisers, publishers and consumers. By the start of 2015 it was clear that even though the product was years ahead of the market, the larger industry was undergoing a sea change and the business needed to evolve. A pure advertising technology company has little value today and cannot survive in the increasingly sophisticated and globally competitive world of digital marketing over the next decade.

Industry Shift

Marketing and Advertising

There's a clear bifurcation in the digital advertising market between so-called “martech” and “adtech” solutions but the definitions are unclear and incorrectly tied to billing models. The real definition is much simpler and comes down to a straightforward comparison of marketing vs advertising.

Modern marketing describes the communication and exchange of information and value between a company and consumers, classically outlined by the “4P's” (product, price, place, promotion). There are newer customer-centric variances today, but the fundamentals remain the same: create or improve a product, define the price and positioning, tell the market about it, produce sales and feedback, and then repeat the cycle.

Under this model, advertising is just a subset of marketing. It's the “promotion” phase of reaching consumers with a message (regardless of the nuances of format, direction or signaling), and this distinction naturally appears in the evolution of adtech's primary focus on distribution while martech handles many of the backend processes involved before and after campaign execution. In essence, all adtech is martech, but not all martech is adtech as it covers a much wider range of concepts.

Due to the relatively complicated nature of accurate and scalable distribution online, adtech has received most of the innovation, investment, and growth over the last few years, but this landscape is quickly changing and will be drastically different over the next decade.

Dissolution of AdTech

Layers of adtech by the flow of money:

Client --> Agencies --> Trading Desk / DSP --> Exchange/Network --> SSP --> Publisher

Adtech is now a commodity as a direct result of all this growth. Ad serving is effectively free, programmatic access has broken down network models, format standardization has deprecated specialist vendors, and business model pressures with inventory oversupply have driven down margins and unit costs.

Although client budgets are increasing, much of the spend is captured by first-party social platforms that offer much better targeting, reach and ad experiences. Many companies are now increasingly unprofitable until they hit scale, if ever, while being fueled by VC money which is quickly drying up from lack of returns and increased competition without differentiation.

The Fall of Networks

Ad networks are predicated on having some unique advantage (like a content niche or audience data), leveraged with scale, to provide valuable inventory for buyers, control pricing for publishers, and enjoy high margins. Five years ago this was a lucrative model but it's no longer viable today..

The rise of programmatic bidding has allowed for precise audience targeting across the web while simultaneously making it easy to put together the perfect niche of sites for every campaign, all without relying on (or paying) a middleman network.

Attention itself, the bedrock of advertising, is now being transferred to social media and digital messaging platforms that live inside walled gardens with their own vertical ad offerings and marketplaces, leaving even less opportunity for traditional networks.

Sell-Side Migration

The last remaining unique offering for ad networks is real estate. Owning the exclusive rights to media not available anywhere else is still an opportunity (and also the primary strength of walled garden platforms).

This shift is clear in the transition of many ad networks into supply-side platforms. Vendors are turning into “software” platforms for publishers and preparing for SaaS pricing models (with some completely giving up media costs) while building out programmatic access. Given the chase after scale and volume, for many networks this is the only viable path.

SSPs provide value by combining several demand sources into optimized revenue output for publishers, which allows them to downsize or eliminate sales operations and reduce income variability. There's also inherent safety and market value in being the final arbiter of access, with incumbent installations and procurement relationships at major media companies.

But this is a dead-end. RTB standardization means anyone can run an auction and the only real specialty is in the formats available. All of the major exchanges are rapidly building better sell-side systems that are more powerful, efficient, cheaper, and can consolidate multiple formats (which just need to be “good enough”). Google's DFP, the gold standard, is also progressing quickly and already contains a unified auction model with default demand through AdX, all of which leaves little reason for any 3rd-party vertical SSP to exist.

Another major problem is that even though revenue flows through an SSP to a publisher, the SSP must treat publishers as their primary clients and the most important relationship to manage. With media owners under constantly growing pressure to grow revenue, they are the worst party to ask for fees (if ad serving fees aren't already negotiated away) and tend to have little loyalty beyond looking for the biggest paycheck. This is a fundamental problem that will stall growth in the short term until the reason for existence for independent SSPs slowly evaporates.

Buy-Side Arbitrage

Some ad networks have carved out a niche by selling their network combined with tools and integrations to media buyers. This is capturing an opportunity between the standardized DSPs and custom bidders that don't have this extra functionality.

This is still under constant pressure from ad exchanges and DSPs consolidating infrastructure and new developments like header bidding that provide faster integrations and more control to publishers. The major buy-side tools are already well developed and will only get more sophisticated, however as long as integrations are a secondary priority, there is a small market space here for these companies.

Realignment and Reintegration

The adtech focus on distribution alone without a strong ROI component has produced an environment that's only concerned with technical metrics like impressions and clicks driven solely by volume. This has resulted in major supply chain issues with billions wasted through ad fraud, bad inventory, poor user experiences, ad blocking and negative brand exposure.

Any measures to combat this are also hindered by the very same programmatic standards that led to such market growth. The latest real-time bidding standards barely support viewability and even transparent and accurate traffic reporting is rare.

Marketing solutions meanwhile are getting much more sophisticated with advances in big data and API integrations leading to very automated and intelligent segmentation, testing, lead gen and post-sales analysis. Recent advancements can been seen in the rise of account-based marketing and the development of marketing clouds from Salesforce, Adobe, Oracle, IBM, SAP, and others to create software stacks that can be used for all marketing functions.

The need to react faster and across more inputs and channels means the slow manual switch from campaign strategy to execution is outdated. Proper distribution and automatic campaign management will be built into martech tools to streamline the entire process with much greater benefits in the usage and collection of proprietary data. The divide between martech and adtech will disappear, and adtech will just be a feature of the martech complex - just as advertising is a subset of the greater marketing concept.

New Opportunities

Building Value

This industry shift is creating a much bigger total market for new companies than the entirety of adtech. Distribution was a hard but narrow problem while the logistics of marketing are much more complex and open to automation and intelligence to derive measurable bottom line value.

Since the sell-side is a dead-end, the only way to grow is to provide more value to marketers who provide revenue for the entire industry and always have been the true clients, and value in a business context is best measured in ROI.

There have already been a handful of companies starting to look for ways to prove real revenue results and a handful of companies are even selling primarily on these business objectives. This is likely where many future successful companies will end up and it not only encourages greater competition and differentiation but decreases much of the inefficiencies and waste of the current model. Given the wide amount of strategies and data available, the market is wide open here for new outstanding players.

Cost Transparency

Media billing is a murky mess with much of it being a cost-plus business that incentivizes the wrong behavior and does not deliver clear value to marketers. With the rise of automation and platforms, most companies are really selling technology or a network with flat costs, and this conflicts with a percentage structure that doubles the price of a campaign for running twice the impressions while having effectively zero marginal cost.

Marketers want more transparency into pricing and if they're purchasing technology, they would rather have it delivered and priced under typical SaaS licensing terms. This brings control back to the marketer with simpler accounting and flexible contracts. Expectations become in line with each service offering - a content management system will be judged on productivity improvements and how it helps marketing efforts, rather than current vendors being seen as an obscure tax on media working dollars.

SaaS models also abstract away the media involvement and are seen as more resources and tools to drive revenue rather than a cost center which improves corporate buy-in and makes ROI and TCO calculations much easier.

Content First

Great content is still the best way to generate brand awareness and inbound interest. Traditional ad messages are increasingly ignored by newer generations, removed altogether by adblockers and turned away by publishers looking for better and faster ad experiences that naturally fit into their editorial stories.

The last few years have borne this out with the rise of social channels, corporate experiences and content marketing by everyone from major companies to local coffee shops. The method works well, but it's also very hard to manage because of the numerous formats and styles that need to be ingested as well as prepared for output to any other source. Search, categorization, translation, workflows and approvals, A/B testing, and more are further challenges that need to be solved with different solutions for each company and vertical.

Content isn't going anywhere but the problems in handling are increasing. A proper solution to these problems creates real baseline value while multiplying the effects throughout the marketing chain.

Distribution

Even as distribution becomes a feature, the main channels used today are email and social. Email requires explicit permission with deliverability and tracking issues while social requires paid promotion in a very saturated and uncontrolled environment. Both formats are also limited by what they can carry and are often promotional messages rather than engaging content.

Trying to syndicate content using these channels is similar to using content recommendation widgets to acquire traffic and ultimately fails for the same reasons; the limited format doesn't actually contain content but relies on asking, begging or even tricking the user, using links to get them to other landing pages and portals.

True distribution (as it works in every other medium) is all about delivering the message (in this case, the content) directly to the user where they are. This has always been possible through manual executions but a fast, seamless and automated solution can vastly expand this space and provide benefits for content marketers and consumers alike.

The Instinctive Solution

The Business

Taking advantage of the new opportunities means evolving as a company and product into a holistic marketing technology provider. Primarily the company will offer a multi-layered platform designed around content marketing and sold on an annual basis.

Individual components are described below but the overarching goal of the software stack is to provide marketers with the ability to manage the entire content lifecycle from creation to distribution to measurement and continuous feedback. This is directed especially to brands themselves but can also be used by larger agencies to augment their content offerings. Beyond the core product, there will be support plans (almost always included in the main license), onboarding fees, and training as the most common package sold to larger clients.

Effectively we will be entering the enterprise sales space with B2B software sold on SaaS license terms. Major obstacles will be procurement, sales & support, and customization to client needs.These obstacles are also strengths that create defenses against competitors as enterprise sales is very much driven by incumbents and process with sales results usually leading to lucrative multi-year deals.

Another way to look at value is to ask what happens if the client cancelled 100% of its advertising budget: would the product still hold value and earn investment? The answer is yes. By fundamentally empowering the client's team instead of being tied to a small niche of output, value is easily measured and constantly delivered under any circumstance.

The Stack

There are 4 main layers that power an efficient content-fueled and consumer-centric marketing system: the content repository, distribution, measurement and integration. Implementations described below:

Content Repository

All content efforts need to start from a base database of available content that the brand has at its disposal. This is implemented as a database that can hold many different types of content assets, in their final assembled form. This is more advanced than digital assets management and is more similar to modern CMS systems user by publishers (who specialize in content management). This repository will act as the foundation of everything else. Brands can input content directly into the database or it can automatically consume for existing sources like company blogs and social channels, so that there are no changes in workflow.

The repository will also handle enhanced functionality including serving assets as products in a ecommerce storefront or survey forms to collect user input.

  • Database to contain all content in a single location with search and tags.
  • Automatically ingest new content from any source to aid in aggregating existing content and letting brands continue existing workflows.
  • Store both atomic assets (images, text, links) and assembled pieces (full articles, video playlists)
  • Assets can include any metadata and represent complex business logic to behave as products for ecommerce, settings under user accounts, or any other richer combination.

Distribution

Content is useless without distribution but due to different definitions of this term, this will encompass both sourcing traffic to hosted content portals and syndicating content directly to existing publishers. This gives us coverage for both without any analytics blind spots and can be used as a transition path for clients wanting to shift to true syndication only or as a continual ongoing experiment for every campaign.

Syndication will also include bringing advanced features like ecommerce checkout directly to the user's experience on foreign sites across the web and mobile. Pricing will remain on a media-cost basis but the lack of required margin means much lower unit costs with more working-dollars for brands for effective lock-in combined with the ease of a single platform.

  • Use publisher network to distribute content directly to users across web.
  • Use standard native units to receive traffic to internal or hosted portal sites.
  • Billing flexibility from vCPM to time-based or flat-fee recurring levels to emulate SaaS charges.

Measurement / Analytics

Measurement mainly serves as a sink for all engagement events that come from the various consumer experiences and touchpoints. This will be used for classic reporting as well as an ongoing corpus and training set to inform, train and optimize machine intelligence to be used in the rest of the stack.

Private and first-party data will also be collected and handled on behalf of the client with various federated security models and procedures including DMP and third-party data companies.

  • Legacy metrics including impressions, clicks, viewability.
  • Content-specific metrics like time spent, view length, scroll depth, active* time.
  • Post campaign metrics for conversions including multi-step funnel, with richer analytics if user is on a hosted portal or hosted distribution format.
  • First-party consumer data collection.
  • Insights derived from collected metrics with descriptive and predictive analytics to offer feedback for future usage.

Integration

There are too many services and tech stacks in use by major corporations and data must be easily available and moved across to where it's needed. All 3 previous layers will be built to serve as the foundational system of record for each module or simply act as a proxy so that clients are free to use whatever system they wish in a configuration that works best for their workflow.

Reducing friction with existing systems will actually increase usage of the main product as it slowly subsumes greater functionality in a single stack while offering the ability to use point solutions where necessary.

  • CRM integration for funnel status combined with content campaign exposure.
  • Post sales analytics on creating new campaigns.
  • Audience retargeting based on funnel performance and CRM attributes.

Services

  • Onboarding and integration - getting accounts started and customizing structure, setup, and connections to infrastructure.
  • Technical support - ongoing customization, setup and connections modification.
  • Team training - continuous offerings for client teams to learn how to use the platform.
  • University / Workshop courses - introduction, best practices, supporting knowledge and in-depth use cases in content marketing, branded content and how to use the Instinctive platform to accomplish goals.

References

  1. http://digiday.com/platforms/ad-tech-consolidates-new-business-models-rise-surface/
  2. http://adexchanger.com/ad-exchange-news/ad-tech-funding-far-dead-lights-guys/
  3. http://www.prnewswire.com/news-releases/sharethrough-reports-significant-growth-in-usage-of-native-advertising-software-by-publishers-in-2015-300193937.html
  4. http://adexchanger.com/platforms/adsnative-raises-8-5-million-series-a-to-take-on-dfp-ad-server/
  5. https://en.wikipedia.org/wiki/Marketing_mix
  6. http://www.mediapost.com/publications/article/266672/why-ad-tech-should-worry-about-the-rise-of-marketi.html
  7. https://www.youtube.com/watch?v=4Tpj8CjYZgA
  8. http://digiday.com/platforms/saas-wars-brands-vendors-prefer-licensing-model-agencies-lag/
  9. http://www.thedrum.com/opinion/2017/02/27/will-february-2017-go-down-the-month-destroyed-adtech