NordVPN Promotion

Home / Blogs

Where Does the Trusted Notifier Fit? Four Abuse Mitigation Models and the Glue Between Them

Seven Months In: Where Does TNN Fit?

We launched the Trusted Notifier Network (TNN) in March 2026, and in the seven months since, the support we have received from the community has been tremendous. Before ICANN’s next round of the New gTLD Program launched—which, understandably, took priority for most of us in the domain community—we were seeing around two to three intermediaries sign up every month.

Those who have not yet signed up are, for the most part, optimistically observing our development and trying to understand how we fit into the larger ecosystem. It’s a question I have been asked many times, and it’s one that has been on top of my own mind as well.

So, what exactly is TNN? It is a new concept—or rather, an old concept with an innovative implementation. Trusted notifier and trusted flagger arrangements have been around for years, but TNN was built from practical need, in an attempt to resolve some of the fundamental issues those arrangements have never addressed. In this article, I’d like to clarify how we see TNN fitting into, and working with, the rest of the abuse mitigation ecosystem.

I believe the industry has been debating tools, when what we are really dealing with are four different models of abuse mitigation, each answering a different question. All four are necessary, but only three of them have been seriously built so far.

1. The Top-Down Model

The first model is regulation, and the EU’s Digital Services Act (DSA) is probably the most developed example, with its statutory “trusted flaggers” designated by national Digital Services Coordinators under Article 22.[1] Similar national regimes are emerging across Asia Pacific.

There is no denying that this model is effective. Legal obligation—with deadlines, audits and penalties attached—moves intermediaries faster than any voluntary best practice document ever has.

However, its limit is geography. A statute stops at the border, but abuse does not. A phishing site targeting Australian bank customers may well sit on a domain registered in North America, hosted in Europe, with its DNS somewhere in Asia, and no single regulator can compel every link in that chain. When cross-border cooperation fails, the only lever a national authority has left is at its own edge, which is blocking. Pushed far enough, the top-down model will inevitably lead to national firewalls—something the internet community should actively work together to avoid, as powers built for phishing seldom stay confined to phishing.

2. The Data Exchange Model

The second model is the exchange of abuse signals, free or paid, through threat feeds, blocklists and signal-sharing platforms. I’m a strong supporter of this model; without it, every intermediary would be fighting blind, and much of the speed at which campaigns are spotted today is owed to these exchanges.

But what does an intermediary actually do with the data? A signal tells you that something was observed—it does not tell you whether it’s actionable against your customer, under your terms of service and in your jurisdiction. The cost of evaluating and acting on that signal remains with the user of the data. Furthermore, data quality and the criteria for why a case is considered abusive are not aligned, often even within the same exchange, which makes it hard to distinguish right from wrong.

As a result, the benefits are unevenly distributed. Large platforms and registries with extensive resources can run the data through their own models and get enormous value out of it. Small and medium-sized intermediaries—often exactly where abuse moves to once the larger players tighten up—frequently lack the resources to even use the data.

3. The Evidence Model

The third model deserves particular credit. Rather than passing raw signals downstream, it uses extensive technical capability to capture and evaluate evidence for each individual case, scores its validity and severity against a published policy, and delivers a standardised report that an intermediary can act on. For abuse types that need specialist expertise, such as illegal pharmaceuticals, specialist verification partners go a step further, adding expert human confirmation to the technical evaluation and setting the evidentiary standard for that category.

This model does something the first two cannot: it replaces subjective judgement with a consistent, objective and highly accurate case-by-case evaluation. More importantly, it creates economies of scale. The heavy technical work is done once and centrally, so a small registrar/host with a handful of staff can receive the same quality of analysis as a global platform—for small and medium-sized intermediaries, this changes the game. The evidence model answers whether a domain is abusive with a rigour the industry has never had before. However, the evidence model has its own challenges: it needs better adoption to scale both its data and its impact. More fundamentally, the interpretation of abuse is not common across all intermediaries, and neither are their evidentiary requirements—what one registrar accepts as sufficient, another may not. The industry needs to work towards a better common understanding of abuse beyond DNS Abuse, and more unified evidentiary requirements. Even then, there will always be grey areas, especially as we move towards abuse types that are less clearly defined, and this is where the interest alignment model will complement it in the future.

4. The Interest Alignment Model

This is the gap that none of the other models attempt to bridge. I believe the root of the distrust between notifiers and intermediaries is not a lack of data or evidence, but misaligned interests.

Consider the three parties involved in abuse mitigation. The intermediary wants to protect its own customers against unfounded complaints and remove only legitimate abuse. The victim, whether a bank, a government agency or a brand, wants to see results from what it spends on protection, and to be able to justify that spend if challenged. The notifier, on the other hand, is paid to satisfy its client, the victim. So, if a notifier detects 1,000 cases and suspects 100 of them are weak, the rational move is to submit all 1,000, as there is no consequence for the 100—the cost of sorting them out simply lands on the intermediary. This is what I call unfair cost transfer, and it is this misalignment that causes the distrust.

How do we realign these interests? I believe it takes two things: accountability and reward.

  • Accountability. There needs to be a meaningful negative impact—one that outweighs whatever can be gained by gaming the system and pushing through a few false positives. In TNN, every notice must be verified with the affected party or victim, and accountability is enforced through a strike-based ledger, where audit gaps, failed victim verification or substantiated inaccuracy reports from intermediaries add up to reduced submission volumes and, eventually, disqualification. This is backed by a chain of contractual indemnity from the victim to the Trusted Notifier, to TNN and on to the intermediary, much like a registrant indemnifies its registrar, the registrar its registry, and the registry ICANN.[2]
  • Reward. Good behaviour needs to be rewarded. It’s part of human psychology: punishment alone rarely builds good habits. In TNN, qualified notices receive a faster review channel, consistent with the principle behind the DSA’s trusted flaggers, and the credential is recognised by every participating intermediary.

Once accountability and reward are in place, interests begin to align again:

  • The intermediary sees its operating cost and legal risk reduced through accountability and indemnity, and receives these notices free of charge.
  • The victim is rewarded with faster review, and can therefore show ROI through the fast-track channel. Its best interest becomes ensuring its provider remains a Trusted Notifier, or it loses the fast track.
  • The Trusted Notifier’s best interest is to remain within the system, as disqualification means losing not just one client but its ENTIRE client base, while staying in is rewarded with a faster review channel that serves as a competitive advantage.

This reverses the commercial best interest: for every party in the chain, the rational choice is now to submit, and act on, only well-founded notices. Long-term trust can only be built where interests align, and this is where it begins.

How the Four Models Work Together

None of these models can win on its own, as each answers only one question, and an intermediary deciding whether to suspend a domain needs all four answered.

  • Top-down (authority schemes) asks “Is this unlawful in our jurisdiction?” It creates urgency and obligation, but it stops at the border and ends in national blocking, with a risk of misuse of power.
  • Data exchange through signal feeds asks “Which domains in my zone look malicious?” It brings speed and breadth of visibility, but the evaluation cost stays with the user, and quality and criteria are not aligned.
  • Data exchange through reporting hubs asks “Is this report complete, and in a standard format?” It produces complete, standardised reports from anyone, but there is no accreditation or accountability behind the reporter.
  • Evidence, through abuse management platforms, asks “Is this threat technically real?” It brings objective, per-case accuracy at scale, but it needs adoption to scale, evidentiary requirements are not unified, and grey areas remain.
  • Interest alignment asks “Who stands behind this notice, and are they accountable?” It brings accountability, reward and indemnity, and its strength depends on how many intermediaries join.

Top-down: a necessary evil. From a multistakeholder purist’s perspective, regulation is a necessary evil—but it is necessary. Change rarely happens without pressure, be it financial, social or political, and regulation is often what drives it. The question is how we channel that pressure. As a non-profit, TNN can induct law enforcement agencies and competent authorities into a bottom-up, multistakeholder process, where their requests are qualified and routed with procedural safeguards rather than imposed through blocking orders. Equally important is working with authorities to encourage intermediaries to adopt trusted channels, which I believe will do far more good than new powers of compulsion.

Data exchange: a trusted lane. TNN doesn’t compete with data exchange; it opens a trusted channel within it. Among the vast volume of signals, a subset can arrive from qualified notifiers, victim-verified and indemnified, which an intermediary can act on at a fraction of the evaluation cost, freeing up scarce analyst time for the rest. For smaller operators, this may well be the only part of the data they can realistically use.

Evidence and interest alignment: two sides of the same coin. This is where I see the closest relationship. The evidence model asks whether the threat is real, while the interest alignment model asks whether the reporter is trustworthy and accountable, and an intermediary needs a yes to both before it can act quickly and safely. The two should coexist and cooperate tightly, as each makes the other stronger. On the one hand, accountability without evidence becomes subjective—a strike ledger is only as fair as the yardstick used to measure notices, and objective evidence scoring is the best yardstick we have. On the other hand, evidence without accountability is like a referee with no disciplinary committee: it can call a foul on every bad report, but it can’t stop the same reporter from sending the next thousand.

Everyone benefits from this. Intermediaries receive notices that are verified, technically scored and legally backed, without paying twice. Evidence providers receive cleaner input at a lower cost per case, good notifiers gain a credential that is objectively demonstrated rather than self-declared, and victims see faster resolution. Most importantly, the cost is finally passed back to the abuser, as attacks become short-lived and no longer yield a return.

If you already use a reporting hub or abuse management platform. Nothing about your current workflow needs to change. TNN notices will be tagged within the platforms intermediaries already use, so they arrive in the same queue, marked as coming from an accountable Trusted Notifier. What a tag alone cannot carry is TNN’s priority commitment and indemnity: the chain of indemnity runs through signed agreements, link by link, so those apply once an intermediary signs TNN’s MoU, free of charge. Nor does TNN pull reports away from the tools you rely on. Where you already use a platform, TNN notices can be delivered through it, and its evidence checks stay exactly where they are. Conclusion: The Glue That Holds It Together

If there’s one thing I’m sure of, it’s that these four models are not rivals. Regulation brings the pressure, data exchange brings visibility and the evidence model brings proof—what has been missing is the glue that binds them together, a reason for every actor in the chain to want the same outcome. That is the role of the interest alignment model: it turns top-down pressure into a community-driven mechanism, gives the data a trusted lane, and gives objective evidence consequences that actually change behaviour. TNN works alongside the existing infrastructure, connecting its parts rather than replacing any of them.

So, where do we go from here? I have three suggestions:

  1. Intermediaries already using evidence-based reporting can add a trusted channel at no cost and see how their risk exposure changes.
  2. Notifiers confident in the quality of their work can put that confidence on the line—and be rewarded for it.
  3. Regulators can ask one simple question of every abuse framework they come across: when a notice turns out to be wrong, who pays? Under most frameworks today, it is still the intermediary that acted in good faith, and that is the cost we need to move back to where it belongs.

There is still work to do on all three, and I would welcome the conversation.

By Alban Kwan, Co-Founder & CEO, Trusted Notifier Network

Filed Under

Comments

  Notify me of follow-up comments

We encourage you to post comments and engage in discussions that advance this post through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can report it using the link at the end of each comment. Views expressed in the comments do not represent those of CircleID. For more information on our comment policy, see Codes of Conduct.

CircleID Newsletter The Weekly Wrap

More and more professionals are choosing to publish critical posts on CircleID from all corners of the Internet industry. If you find it hard to keep up daily, consider subscribing to our weekly digest. We will provide you a convenient summary report once a week sent directly to your inbox. It's a quick and easy read.

Related

Topics

New TLDs

Sponsored byRadix

Domain Names

Sponsored byVerisign

Brand Protection

Sponsored byCSC

DNS Security

Sponsored byWhoisXML API

Cybersecurity

Sponsored byVerisign

DNS

Sponsored byDNIB.com

IPv4 Markets

Sponsored byIPv4.Global

NordVPN Promotion