Basic Attention Token logo Basic Attention Token

BAT Roadmap 1.0

by Brendan Eich | June 16, 2017 | General , Token


I cofounded Mozilla and wrote its technology roadmaps back when we were rebooting the Netscape-contributed browser code via open source. Here’s the first one. This one from 2003, by David Hyatt and me, charted the direct course to Firefox.

These documents set direction and listed goals and requirements. Equally important, they renounced non-goals and listed what was off the agenda. For the BAT platform, any tracking or targeting from remote ad servers or exchanges is excluded by construction.

The map is not the terrain, so each roadmap ages past a point of usefulness and becomes a relic. I expect the same to happen with the BAT project’s roadmaps. This roadmap is likely to cover the next year to 18 months.


The BAT sale provides the development team with enough resources to complete the platform work we started last fall using bitcoin in the “Brave Payments [beta]” experimental automated micro-donation system already in Brave. We intend to use the funds as follows:

  1. Recruit and retain Ethereum, machine learning, devops, project management, ad sales, UX design, and more browser development people, to build out the BAT platform in the stages outlined below.
  2. As the BAT platform develops and gets integrated with Brave, evangelize other attention-economy apps, from browsers to games and beyond, to join an organization we will endow to promote user-first, private, and anonymous ads, donations, micro-paywall-piercing “link gifting”, and other token-based premium products, and share the BAT among apps other than Brave. We aim to make this a non-profit.
  3. Create a source of seed funding, with exact structure to be determined, to sponsor entrepreneurs to create new BAT apps. This may take the form of an independently judged prize system in which apps must follow the endpoint software rules and on-chain smart contracts prototyped by the development team (1) and codified by the member organization (2).
  4. Fund greater user growth of Brave and other BAT apps by rewarding users who choose to opt into the BAT platform, to get to greater scale sooner with the brands and agencies buying user-private ad slots, and (where publishers partner with us) ad slots on pages. Growth requires promotion and distribution above organic rate and scale, so we will fund it with some of the sale proceeds.

The BAT platform work must be staged to fit modern “agile”/non-waterfall software development methods. In particular, it must fit Brave’s six-week chromium update cycle and the finer-grained milestone release structure that we use to manage scope creep and increase predictability.

The program of work outlined in this roadmap uses code-names borrowed from the U.S. Space Program. Reforming the attention economy of ad-tech and payments may not be rocket science, but it is still ambitious. We need to manage scope creep at a larger scale than the Brave development cadence.



BAT Mercury (summer 2017)

  • BAT wallet integrated into Brave
  • Convert Brave Payments from bitcoin to BAT
  • Extend Brave’s secret-key sync to include the Brave ledger and wallet
  • Compile demand-side dashboard data into BAT ad and offer catalog that is downloaded and updated on-device; scraped and mock ads/offers with micro-BAT subsidized revenue
  • Develop the on-device machine learning models that privately match catalogued ads and place them in user-private slots based on user intent signals, from search queries (you own your query log, Google does not) to surfing, to researching, to purchasing
  • Design research and engineering development of user-private ad slot form factors
  • Measure ad engagement and user satisfaction via Basic Attention Metric (BAM) system
  • Extend the Brave Ledger ANONIZE ZKP-based proof protocol (Zero Knowledge Proof) to convey ad attributions and confirmations as well as transactions

BAT Gemini (fall 2017/winter 2018)

  • KYC (humint + machine learning), rate, and flow control for anti-fraud
  • User Growth Pool (UGP) grants to users who opt into the new model
  • Further machine learning development and optimization
  • Further demand-side dashboard to BAT ad/offer catalog work
  • Ledger- and Ethereum-attested dashboard analytics for early publishers interested in BAT indirect ad partnership trials
  • User-private ad trials with select agency and other demand-side / lead-gen partners
  • One premier publisher partner to co-develop publisher-provided inventory system
  • Revenue as soon as system performs as well as or better than status quo ad-tech

BAT Apollo (rest of 2018)

  • Real ad revenue that scales with user growth; further UGP grants & growth hacks
  • Donation flow fee revenue, smaller but scaled as far as possible via incentives to users who get BAT revenue from ads and want to give back to their pinned and top sites
  • Major work to move from Brave Ledger confirmation and revenue flows to entirely decentralized on Ethereum flows using a state channel with ZKP for anonymity
  • Evolved BAM options based on premium attention models — price discovery at high end
  • Aggregated, large-anonymity-set reports and forecasts on the Ethereum blockchain, with competitive demand (advertiser) and supply (publisher) dashboards for performance measurement, optimization and sales planning
  • BAT integration into other apps based on open source & specs for greater ad buying leverage and publisher onboarding



Some of this work will happen ahead of the rough dates in parentheses above, enabling other work to happen sooner. Other work that’s hard to estimate right now may take longer. We will adjust the roadmap quarterly and build a new one when this one has fulfilled its purpose.

Much depends on our hiring quickly and well. With the right talent engaged and working well together, I am confident we can build and scale this system to have complete user control and much better economics than the failing ad-tech ecosystem upon which the Web relies today. Let’s build it together. We are all-open-source, and we are hiring.