48.2%
Average measured successful-removal rate in one independent user study—not a reason to give up, but a reason to verify.
Evidence S1
Product · Engineering · Market · GTM
The deletion atlas.
A practical plan to build Clean Slate into the trusted command center for reducing a person’s digital footprint—one clear decision, verified outcome, and honest boundary at a time.
01 / 12
Executive thesis
Win on outcomes, not on an inflated site count.
Clean Slate should become the interface between a person and every legitimate path to reduce their online exposure. It will use official rails when they are stronger, automate narrow workflows when safe, and clearly explain what cannot be erased.
The decision
Build a US-first, whole-footprint cleanup product around a free Actionability Map and a paid, continuously verified Proof Ledger.
1
Make the problem feel finite.
Turn a frightening digital footprint into a short queue: automatic, approve once, confirm, guided, or contested.
Product rule
Every result gets both a risk score and an actionability class.
Every result gets both a risk score and an actionability class.
2
Charge for continuity and proof.
Discovery is the acquisition hook. Recurring removal, verification, monitoring, custom cases, and household coordination are the subscription.
Commercial rule
Never paywall basic knowledge of exposure.
Never paywall basic knowledge of exposure.
3
Keep AI in a bounded advisory role.
Models classify, redact, summarize, draft, and prioritize. Deterministic code chooses recipients, disclosures, and execution.
Engineering rule
No model autonomously deletes or transmits identity data.
No model autonomously deletes or transmits identity data.
4
Build the moat from verified operations.
Every completed case improves connector reliability, outcome estimates, jurisdiction routing, and the next user’s plan.
Data rule
Outcome data beats nominal coverage.
Outcome data beats nominal coverage.
02 / 12
Market timing
The rails are improving. The user experience is not.
Regulation, platform tools, and privacy services create more removal paths than ever. The opportunity is orchestration: choose the best path, minimize disclosure, manage follow-up, and prove the result.
9%
of 522 California-registered brokers were fully compliant with transparency requirements in a 2026 audit.
Evidence S2
600+
active brokers can receive one California DROP request, with hashed matching and rolling processing.
Evidence S3
Interpretation
The market is real, but noisy. A user cannot infer success from “requests sent” or “sites covered.” Clean Slate’s durable advantage is a trustworthy state machine—from discovered through independently verified—plus a plain explanation whenever deletion is impossible.
Why the timing is favorable
Official infrastructure is consolidating
California DROP creates a powerful free route. Clean Slate should deep-link, explain, and track it—not recreate it. S3
Search platforms expose useful controls
Google can remove eligible results, but source content may remain. That distinction is a core UX opportunity. S4
Consumer expectations are moving up
Free scans, screenshots, recurring monitoring, custom cases, and concierge support are already category norms in different products.
03 / 12
Market & competition
A crowded category with an unclaimed center.
The leaders each own a useful promise. None fully owns the calm, end-to-end command center for brokers, abandoned accounts, search, social exposure, custom URLs, official rights, and durable proof.
Strategic interpretation, not a quantitative market survey. Positions synthesize current public product claims and the existing Clean Slate research snapshot. See sources.
Competitive readout
| Product | What it owns | Free hook | Opening for Clean Slate | Strategic response |
|---|---|---|---|---|
| Incogni 420+ automated brokers advertised |
Set-and-forget broker automation | Limited core free experience | Less centered on an inspectable whole-footprint plan | Match automation; beat on cross-surface orchestration and proof S7 |
| Optery | Free visibility, screenshots, recurring scanning | Strong exposure report | Result volume and plan complexity can create review work | Make the first scan more decisive, not merely exhaustive S8 |
| DeleteMe | Brand trust and human-managed service | Education + Permission Slip | Core promise remains broker-centric | Compete on product clarity; add human escalation selectively S11S13 |
| Kanary | High-risk attack-surface protection | Community entry tier | Premium and specialist orientation | Serve mainstream transitions first; preserve a concierge path S9 |
| EasyOptOuts | Price/value benchmark | No full free service | Deliberately narrow scope | Do not race to the bottom; prove broader value S10 |
| Permission Slip / YDR | Simple legal-rights actions | Useful free workflows | Follow-up and verification still require work | Use as benchmarks—and route to them where appropriate S11S12 |
| Clean Slate proposed target |
Whole-footprint command center | Actionability Map | Must earn trust and connector reliability from zero | Proof, least disclosure, honest boundaries |
04 / 12
Customer wedge
Start with a moment of urgency, not a generic privacy persona.
“Privacy-conscious adult” is too broad to acquire efficiently. The launch wedge is a person facing a concrete transition or elevated exposure who needs a credible plan today and continued monitoring afterward.
Life-transition cleaner
A job search, move, breakup, new child, name change, or identity-theft event turns background unease into a deadline.
- Show me where I appear without making me learn privacy law.
- Tell me what can disappear and how much work it takes.
- Handle the repeatable work; ask me only for key decisions.
- Give me evidence I can trust and revisit.
Elevated-exposure professional
Founders, creators, executives, healthcare workers, public servants, journalists, and activists need broader monitoring and escalation.
- Reduce doxxing and impersonation surface.
- Protect household members and aliases.
- Escalate hard cases to a trained human.
The entry promise
Free wedge
“In one private scan, know what can disappear, what needs you, and what has to stay.”
05 / 12
Product system
The interface is a queue of solvable work.
Clean Slate’s core interaction is not a fear score or an endless list. It is an Actionability Map that translates exposure into the smallest safe next step.
A1
AutomaticAuthorized, reversible, standardized request. The system can submit and verify it.
No action
A2
Approve oncePrepared batch shows exact recipients, legal basis, and fields before submission.
One approval
A3
ConfirmUser must click an email, sign in, solve a CAPTCHA, or approve an irreversible step.
Short assist
A4
GuidedThe platform requires direct action; Clean Slate opens the right setting and tracks completion.
User-led
A5
ContestedPublic record, lawful publication, legal exception, or refusal. Offer de-indexing, correction, hardening, or monitoring.
Explain options
Experience sequence
01
Minimal identity
Name, region, optional aliases and identifiers—only enough to begin.
Trust moment
Explain purpose and retention field by field.
Explain purpose and retention field by field.
02
Private discovery
Public search, broker queries, user-supplied URLs, and optional local-first account evidence.
Activation
First useful result in under three minutes.
First useful result in under three minutes.
03
Confirm matches
Resolve ambiguity with evidence, not model confidence alone.
Quality
Optimize precision before result volume.
Optimize precision before result volume.
04
Approve a plan
Batch safe work; isolate sensitive disclosures and irreversible decisions.
Conversion
Paid value is visible before checkout.
Paid value is visible before checkout.
05
Verify and monitor
Capture receipts, broker responses, before/after evidence, deadlines, and reappearance.
Retention
The ledger becomes the reason to stay.
The ledger becomes the reason to stay.
06 / 12
Business model
Monetize relief, continuity, and hard cases.
The free product establishes trust and exposes the plan. The paid product removes recurring labor, verifies outcomes, and catches reappearance. Premium tiers add people and human expertise—not fear.
Free · acquisition
$0
Know the shape of the problem.
- Minimal-PII exposure scan
- Confirmed matches and Actionability Map
- Official government and platform routes
- Three guided removals per month
- Local progress export
Success: first confirmed result + first completed action.
Complete · recommended
$99 / year
Or test $12 monthly. Final price follows unit economics.
- Continuous broker and public-web monitoring
- Managed recurring broker removals
- Bulk approval and confirmation inbox
- Search and cached-result workflows
- Ten custom URL cases per month
- Proof Ledger and reappearance alerts
- User-selectable model
Success: verified removals + sustained clean days.
Household / Concierge · later
TBD
Validate demand and labor cost before launch.
- Multiple identities and aliases
- High-risk response mode
- Human-reviewed custom removals
- Regulatory escalation
- Public-presence allowlist
Success: higher ARPU without service-margin collapse.
Pricing logic
The $99 annual hypothesis sits well above the narrow $19.99/year EasyOptOuts benchmark and close to Incogni’s captured annual entry price, while packaging a broader job. It is a recommendation—not a validated willingness-to-pay result. S7S10
Model choice as trust control
Expose an approved catalog: Fast, Balanced, Most capable, and Bring your own key. Translate model differences into privacy, cost, and quality—not provider fandom.
07 / 12
Engineering plan
Treat every deletion as a durable, auditable workflow.
The product should be assembled as a privacy-safe orchestration system: an encrypted identity vault, a versioned connector registry, durable jobs, an append-only evidence ledger, and an AI sidecar that never owns execution.
Identity vault
Field-level encryption, isolated keys, region-aware storage, explicit retention, and a one-control export-and-delete flow.
Targeted disclosure
Each connector declares required fields. The approval screen shows exactly what will be sent and why.
Durable workflows
Retries, confirmation windows, statutory deadlines, idempotency keys, and terminal reason codes are first-class state.
Append-only proof
Outbound payload hash, response, receipt, before/after evidence, verifier version, and reappearance checks are preserved.
Local-first account discovery
Prefer browser or on-device extraction of account evidence. If OAuth is necessary, request narrow scopes and discard message content.
Bounded AI
Models never select the recipient, determine legal authorization, merge raw PII, or trigger irreversible execution.
Minimum connector contract
Core data model
| Entity | Purpose | Sensitive boundary |
|---|---|---|
| IdentityProfile / Identifier | Names, aliases, regions, emails, phones, addresses | Vault; field-encrypted; short retention where possible |
| Exposure / MatchEvidence | Discovered item plus why it is likely the user | Redacted preview outside vault |
| Request / RequestEvent | Durable state machine and append-only history | Payload hash, not raw payload, in general logs |
| Connector | Required fields, policy, execution, verification | Version-pinned for every request |
| VerificationRun | Independent outcome check and reappearance status | Evidence encrypted at rest |
| ConsentGrant / ModelPolicy | Authority, scope, expiry, model/provider preferences | Immutable consent events + revocation |
08 / 12
Roadmap & operating plan
Prove trust and outcomes before breadth.
The build sequence deliberately puts consent, request state, proof, and verification ahead of a large connector count. Each phase has a measurable gate.
WEEKS 0–6
Foundation
Identity vault, consent model, request state machine, event ledger, connector SDK, manual-ops console, billing shell, and official-rail guides.
Exit gate
- 10 connectors in staging
- 0 raw PII in logs
- Full export + delete tested
WEEKS 6–12
Private beta
25–50 design partners, 30–50 high-value US connectors, public-search discovery, match confirmation, batch approval, verification runs, and high-touch operations.
Exit gate
- ≥70% first-value rate
- ≥1 verified removal/user
- <15 min median user effort
MONTHS 3–6
US general availability
Free Actionability Map, annual Complete plan, 100 account guides, confirmation inbox, custom URLs, reappearance alerts, referrals, and support playbooks.
Exit gate
- Paid retention signal
- Support margin understood
- Connector SLOs stable
MONTHS 6–12
Compounding scale
Local-first account discovery, household profiles, connector health automation, partner APIs, risk-mode workflows, and targeted human escalation.
Exit gate
- Healthy payback
- Reliable proof loop
- Partner channel validated
MONTHS 12–18
Jurisdiction expansion
GDPR workflows, localized rights routing, expanded identity-verification patterns, counsel-reviewed authorized-agent models, and multilingual guidance.
Exit gate
- Local counsel sign-off
- Regional data controls
- No one-size-fits-all forms
Lean launch team
Product / founderStrategy, UX, design partners, partnerships
2 full-stack engineersApp, workflow engine, billing, observability
Privacy / connector engineerRegistry, verification, browser workflows
Privacy operationsExceptions, QA, evidence, support playbooks
Fractional expertsSecurity review, privacy counsel, threat modeling
Recommendation. Add dedicated growth only after activation, proof quality, and retention show repeatability.
09 / 12
Go-to-market
Let the product demonstrate the problem—and the relief.
The growth engine starts with a useful free result, earns a first win, and converts on continuity. Paid acquisition should wait until verified outcomes and annual retention are visible.
Product-led conversion path
01
High-intent visitSEO · referral · partner02
Minimal private scanlow-friction trust03
Confirm identity matchesprecision moment04
Complete first actionactivation05
See paid cleanup planvalue before checkout06
Receive proofretention + referralChannel order
Search
High-intent removal guides, “remove from X,” official-rail explainers, state rights, and abandoned-account checklists.
Product loop
Redacted before/after proof report, privacy checkup reminders, and referral credit after a verified win.
Partners
Identity-theft support, cyber insurance, executive protection, employee benefits, family offices, and legal referrals.
Earned media
Publish a quarterly Outcome Index: verified rates, user effort, broker friction, and reappearance—not vanity coverage.
Paid
Only after funnel and retention evidence; bid on event-driven intent, not broad privacy anxiety.
Launch program
DESIGN PARTNERS
25–50 people, manually observed
Recruit from real transition moments and elevated-exposure groups. Watch every hesitation. Personally resolve failures. Do not turn these users into testimonials without explicit, revocable consent.
Learn
- Trust objections
- Match quality
- Time to relief
PRIVATE BETA
100–250 invited users
Open the free map and annual plan. Publish connector status honestly. Use a waitlist segmented by urgency and geography.
Prove
- Activation
- Conversion
- 30/60-day value
PUBLIC US
Search-led release
Launch with 100 high-demand guides, official-rail tracking, an outcome methodology page, a security explainer, and a transparent limitations page.
Scale
- Organic demand
- Referral loop
- Annual retention
10 / 12
Metrics & economics
Measure relief the way a user experiences it.
The metric system must reward verified outcomes, precision, and low effort. Nominal coverage and requests sent are diagnostic inputs—not success.
North star
Verified exposure reductions per activated user
Count a reduction only after an independent check, with a reason code and evidence timestamp. Track reappearance-free days as the long-term quality measure.
Time to first confirmed exposureDiscovery usefulness
First-action completion rateActivation
Match precisionUser-confirmed correctness
Verified removal rateOutcome quality by connector
Median user minutes / caseFriction removed
Reappearance-free daysDurability
Free → paid conversionPlan value
Annual gross retentionContinuity value
Bottom-up revenue scenario
Simple, inspectable math
At the recommended $99 annual price. This is not a TAM claim.
10,000 customers$0.99M ARR
50,000 customers$4.95M ARR
100,000 customers$9.90M ARR
250,000 customers$24.75M ARR
Operating targets to validate
Gross margin
After verification, connector operations, support, and model usage.
≥70%
Blended CAC
Target supports annual-plan payback without assuming perfect retention.
<$60
Annual gross retention
Proof that monitoring and recurrence create sustained value.
≥70%
High-touch case share
Percentage of paid cases requiring human intervention.
<15%
Connector verification SLO
Supported connectors with a fresh health check.
≥95%
Cost reality
Model tokens are unlikely to be the dominant unit-cost driver. Connector maintenance, verification, exceptions, and human custom-removal work are the variables to instrument from the first beta. Vercel’s zero-markup gateway pricing is helpful, but it does not solve operational cost. S6
11 / 12
Risk register
Trust can fail faster than software.
The highest risks are not ordinary feature bugs. They are wrongful deletion, unnecessary disclosure, false claims of success, regulatory overreach, and a service operation that quietly destroys margins.
R01
Wrong identity / wrongful removalRequire user-confirmed matches for ambiguous records, isolate irreversible actions, preserve public-presence allowlists, and record consent at the action level.
Critical
R02
Privacy service becomes a data honeypotField encryption, separated keys, least disclosure, short retention, redacted logs, access audit, breach response, and one obvious self-delete flow.
Critical
R03
Claims exceed real outcomesPublish methodology, distinguish submitted/acknowledged/removed/verified, disclose connector health, and never turn a missing result into a success by default.
High
R04
Automations break as sites changeVersioned connector contracts, synthetic health checks, feature flags, safe fallbacks, and a maintained queue ranked by user impact.
High
R05
Authorized-agent model is invalidJurisdiction-aware policy engine, counsel review, explicit authority scope, user-executed fallbacks, and no universal legal form.
High
R06
Human operations erase marginMeter exceptions, cap custom cases, price concierge separately, build reusable playbooks, and expose “guided” honestly instead of hiding labor.
Medium
R07
AI leaks or fabricates sensitive detailPlaceholder redaction, allowlisted tasks and models, no autonomous execution, output validation, provider controls, and user-visible model policy.
High
R08
Official tools make paid work redundantRoute users to the best free rail. Monetize discovery, orchestration, cross-surface coverage, proof, monitoring, and exceptional cases.
Strategic
12 / 12
Decisions & next 30 days
Turn the blueprint into evidence.
The next month should reduce the four biggest uncertainties: trust, match quality, connector reliability, and willingness to pay.
DAYS 1–7
Freeze the trust model
Threat model, data inventory, retention schedule, consent scopes, model boundary, and incident owner. Review with privacy counsel and security lead.
Owner · founder + engineering
DAYS 1–10
Recruit 25 design partners
Segment by trigger moment. Run problem interviews and a clickable flow. Test what users will disclose, what proof they trust, and where they expect human help.
Owner · product
DAYS 7–21
Build 10 connector vertical slices
Each slice must include discovery, approval, request, confirmation, verification, failure reason, and reappearance—not only form submission.
Owner · connector engineering
DAYS 14–24
Run a willingness-to-pay test
Test free-only, $99 annual, $12 monthly, and a higher-touch option after showing the actual plan. Measure checkout intent, not polite preference.
Owner · product + growth
DAYS 21–28
Instrument the proof loop
Define outcome reason codes, evidence requirements, verifier cadence, connector SLOs, and dashboards for user effort and high-touch intervention.
Owner · data + engineering
DAY 30
Hold the go / narrow / stop review
Advance only if people trust the workflow, matches are precise, at least one removal can be verified reliably, and the paid plan solves a larger job than official free tools.
Owner · founding team
Evidence register
S1Measuring the Accuracy and Effectiveness of PII Removal ServicesIndependent 2025 study; coverage, match accuracy, and removal outcomes. Source ↗
S2Privacy Without Remedy2026 California broker compliance audit. Source ↗
S3California DROP explainerOfficial scope, hashing, dates, and processing cycle. Source ↗
S4Google Results about youMonitoring and removal scope; source-removal distinction. Source ↗
S5Vercel AI Gateway models & providersUnified API, discovery, provider switching, and fallbacks. Source ↗
S6Vercel AI Gateway pricingZero-markup token pricing and current credit model. Source ↗
S7IncogniCurrent advertised plans, automated broker coverage, and custom removals. Source ↗
S8Optery pricingFree monitoring and paid removal features. Source ↗
S9Kanary 2026 Product BriefAttack-surface positioning and breadth. Source ↗
S10EasyOptOuts pricing$19.99 annual value benchmark and focused coverage. Source ↗
S11Permission SlipFree company requests and paid bulk/concierge layer. Source ↗
S12YourDigitalRights.orgFree, open-source rights-request and escalation service. Source ↗
S13DeleteMe individual broker removalsCaptured current individual offer. Source ↗
S14EU GDPROfficial consolidated regulation; Article 17 erasure right and exceptions. Source ↗
S15Google private information removalEligible PII and the difference between Search removal and source removal. Source ↗
Method and confidence. This document combines a frozen snapshot of the Clean Slate product prototype and internal strategy with official product pages, government guidance, and independent academic research checked on July 28, 2026. Competitor pricing and coverage are volatile point-in-time facts. Market positions, pricing, sequencing, operating targets, and the bottom-up revenue scenarios are explicitly recommendations or models—not reported market facts. No customer testimonials, certifications, proprietary success rates, or market-size claims have been invented. The full machine-readable evidence manifest accompanies this document.