[Inference | Core conclusion] ONAR has purchased an agency business with historical profits and its data workflow. Future value depends on whether, while retaining clients and the team, technology investment can be turned into billable capacity and cover financing, integration and maintenance costs. The most noteworthy constraint comes from the contract: the gross‑profit earn‑out excludes headquarters and internal charges, so there remains a gap that must be verified separately between earn‑out success and the group’s cash return.
Executive summary: four conclusions determine post‑investment priorities
- [Fact] ONAR on September 29, 2026 acquired all equity of Advertise Purple, announced the next day. The fixed consideration consists of USD 12,825,000 cash and USD 7,000,000 seller notes, plus an earn-out of up to USD 8,000,000 gross profit. [S1]
- [Fact] The target disclosed that in 2025, net revenue was about USD 17,100,000, net profit USD 4,400,000, and adjusted EBITDA about USD 6,600,000; historical data unaudited. Adjusted EBITDA includes itemized add-backs of about USD 2,151,000, with the total differing from about USD 6,600,000 due to rounding. [S2]
- [Inference] The gross profit earn-out protects business growth but does not fully constrain group technology costs, headquarters costs, and financing costs. Contract accounting and cash management ledgers should be set up in parallel, and double-counting of data and client contributions should be avoided.
- [Recommendation] Within the first hundred days, prioritize validating two real workflows, using client queues, complete manual review hours, and receivable data for acceptance. Before AI operating results are achieved, subsequent M&A expansion should be assessed separately from system replication capability.
1. Fact Base: Buyer obtains control, original team retains key operating boundaries
[Fact] The acquisition entity is ONAR subsidiary ONAR LLC, making the target an indirect wholly-owned subsidiary. The seller notes carry an annual interest rate of 8% and mature in 36 months, with a guarantee from the parent company; the transaction still involves customary post-closing adjustments. [S1]
[Fact] Advertise Purple's CEO Jonathan Moisan and CGO Rowland Hazard continue to lead the business. Bloom assumes the role of client performance dashboard and internal affiliate marketing operations platform, containing over 111,000,000.00000001 performance records; the company says it has added natural language access capability and plans to combine with ONAR Labs' Retina AI and Cortex. [S2]
[Inference] This acquisition simultaneously obtains client relationships, business teams, software, and historical data. Acquiring these assets does not mean they have formed a unified product. Identity matching, data permissions, metric definitions, client service responsibilities, and financial ownership are specific integration tasks to be completed.
[Addendum] This report builds on the internal ONAR transaction short case, adding contract incentives, historical profit quality, financing constraints, and hundred‑day acceptance analysis. It does not recount the same transaction as a new event, nor assume the buyer has an undisclosed on‑site engineering team.
2. Transaction Economics: First assess cash profit, then discuss AI premium
[Inference | Arithmetic] Using the disclosed approximately USD 6,600,000 adjusted EBITDA, the fixed consideration of USD 19,825,000 is about 3.0 times, and the maximum consideration of USD 27,825,000 is about 4.2 times. This is merely a rough “consideration / historical adjusted EBITDA” multiple, without adjusting for net debt, cash, working capital, the present value of the earn-out, and transaction fees, and therefore cannot be called a standard EV/EBITDA transaction multiple.
[Fact] The company's main add-backs converting net profit to EBITDA include compensation-related USD 712,000, state franchise taxes USD 520,000, website and software development USD 435,000, management compensation USD 144,000, excess rent USD 116,000, and other items of about USD 224,000. The company says these items were identified by an independent profit quality review, but the full review report is not disclosed in these materials. [S2]
[Inference] The above add-backs total USD 2,151,000, representing about 33% of the disclosed adjusted EBITDA. Whether the add-backs are reasonable should be examined item by item to determine if they will still be incurred after closing: software development may be a continuing expense to maintain product competitiveness; the departure of management personnel may require replacement positions; and the ability to terminate the lease determines whether rent can be reduced.
[Recommendation | Cash Bridge] Starting from the verified operating profit, sequentially deduct ongoing software investments, cash taxes, cash interest, increase in working capital, and necessary capital expenditures, then assess the cash available for debt repayment and distribution. Software expenses already deducted in the expense line should not be deducted again, nor should capitalized development costs be completely ignored.
[To be verified] The target company's financial statements, QoE report, personnel and compensation list, classification of development expenditures, and closing working capital adjustments are required. The gap between historical net profit and adjusted EBITDA is insufficient to demonstrate AI value creation.
3. Earn-out design: business expansion incentives may be separated from group profit.
[Fact] The purchase agreement sets three annual earn-out periods ending on 2027, 2028, and September 30, 2029. In the first period, gross profit between 13,500,000 and USD 15,500,000 linearly triggers a payment of 0 to USD 2,000,000. In the second period, gross profit between 19,000,000 and USD 21,000,000 triggers a payment of 0 to USD 3,000,000. In the third period, gross profit between 23,000,000 and USD 25,000,000 triggers a payment of 0 to USD 3,000,000. [S3]
[Fact] Contractual gross profit is calculated after deducting specified direct service costs, excluding headquarters expenses, internal charges, and management fees; revenue from buyer alliance marketing generated specifically by the target’s employees, accounts, or systems is also included. The agreement also requires maintaining separate accounting and restricts business transfers and unjustified termination of key management positions. [S3]
[Inference | Incentive Effect] This arrangement helps limit the buyer’s ability to lower the earn-out through cost sharing or client shifting, and allows the original team to participate in cross‑selling. However, if the central technology team expands, inference costs rise, and product R&D intensifies, contractual gross profit may still be met while the group’s cash contribution does not improve.
[Recommendation] Create two parallel statements. The contract statement strictly follows the agreement and does not alter seller rights through management adjustments; the internal statement records ONAR Labs’ R&D, delivery, customer service, and financing costs based on actual resource consumption. The two are linked by a reconciliation table showing “Earn‑out gross profit — actual business contribution — group cash contribution”.
[Inference | Organizational Impact] Protecting the original team increases the likelihood of retaining customers and expertise, and also means the central team must obtain business adoption through co‑design and verifiable returns. The speed of technology migration should be constrained by customer continuity and data quality, avoiding using system uniformity itself as a performance metric.
[To be verified] The actual accounting implementation of each contract metric, cross‑client revenue attribution, and dispute resolution. The 2025 net revenue metric and the future contract gross profit metric have not been fully reconciled, so they cannot be directly used to calculate the earn‑out growth rate.
4. Data Asset Examination: How 111,000,000.00000001 records become a sellable capability
[Inference] Record count is merely a description of data volume. Whether the data can create an advantage must pass four tests: rights, quality, transferability, and business impact. Owning the platform does not automatically grant the right to use all client data for cross‑client training or marketing.
- Rights and scope: confirm on a case-by-case basis the permitted uses, retention periods, and opt‑out deletion requirements stipulated in client contracts, network agreements, and privacy terms. Clients whose data usage rights are not confirmed should not be included in cross‑client model trials.
- Quality and semantics: verify that transactions, returns, commissions, coupons, and attribution events are de‑duplicated; distinguish transaction amount, agency net revenue, incremental client profit, and platform attribution amount.
- Transferability: set aside test sets across different industries, average order values, sales cycles, and traffic structures to assess whether historical patterns remain effective across clients; avoid mistaking the advantage of head‑client samples for a universal capability.
- Business impact: use the client’s actual contribution profit, verifiable new orders, renewals, or paid expansions as the endpoint. More accurate forecasts and faster dashboards are intermediate metrics.
【Recommendation | Product roadmap】First integrate Bloom’s data traceability and explainable suggestions, then incorporate Retina AI forecasting and Cortex attribution capabilities. Preserve source, rules, model version, manual adjustments, and final execution records at each step. This describes a suggested route, not a completed company product architecture.
【Falsification condition】If the data primarily serves display convenience, cannot be legally used across clients, or the model’s gain comes only from a few clients, valuation should revert to the cash flows of agency services and single‑client software tools, and should not assign an unverified premium to a data platform.
5. Two FDE‑style pilots: from workflow to financial acceptance
【Inference | Suggested organization】Form a team of business leaders, alliance marketing experts, data/application engineers, and financial analysts. Engineers should directly observe tasks, modify the system, and jointly accept with the business, which aligns with the report’s definition of an FDE‑style mechanism. ONAR has not yet disclosed the corresponding configuration.
- Pilot A: partner selection and commission recommendation. Record the full chain of manual search, data cleaning, recommendation generation, approval, execution, and review. Initially run in recommendation mode, with the business approving commission changes; evaluate new client profit, anomalous commissions, effective partner count, and total delivery labor hours.
- Pilot B: client monthly performance review. Link orders, refunds, commissions, and service contracts to generate traceable analysis. Assess the cycle from billing to delivery, financial reconciliation differences, review labor hours, client adoption rate, and renewals—not just the speed of report generation.
- Replication test: a second business team deploys the same components, recording new configurations, manual support, and maintenance labor hours; only if unit deployment cost declines while maintaining the same quality can the roll‑up replication argument be supported.
【Recommendation | Attribution method】Match clients by industry, size, and service complexity, and roll out in batches; record results from the original process versus the AI‑assisted process. Use random grouping when random assignment is possible; otherwise, clearly note selection bias in observational comparisons. Agencies that voluntarily choose the simplest client pilots will overestimate company‑wide benefits.
[Advice | Avoid Double Counting] If saved labor hours are used to take on new customers, only the corresponding incremental contribution should be counted; the same batch of hours cannot also be counted as wage savings. Idle time that does not reduce outsourcing, hiring, or overtime expenses should first be recorded as capacity release.
6. Operational Sensitivity: Customer churn can quickly swallow technology gains.
[Inference | Illustrative Calculation, Not Forecast] Using the target 2025 net revenue of USD 17,100,000 as a fixed size reference, assume net revenue increases by 5% and the marginal contribution rate of new revenue is 50%, yielding an additional contribution of about USD 427,500. If the incremental annual cost for AI, internal engineering and maintenance is USD 500,000, the first-year operating net increment is approximately negative USD 72,500, not yet accounting for one‑time integration, financing and taxes.
[Inference | Another Sensitivity] Under the same 50% marginal contribution assumption, if existing customer revenue declines by 5%, contribution drops by about USD 427,500. Technology savings must first cover this loss. Here, revenue growth or marginal contribution rate is not used as an ONR guideline; both require replacement with customer‑level data.
[Advice | Calculation Framework] Renovation operating net increment = attributable new contribution from existing customers + new contribution from new customers + realizable cost reductions – additional technology and maintenance costs – incremental customer service costs. Cash impact further deducts one‑time integration, working capital, cash taxes and related financing payments.
[Judgment] Improvement in profit per individual customer is a better gauge of renovation quality than consolidated revenue growth. New acquisitions, price increases, customer mix changes and advertising market shifts should be listed separately to avoid folding all changes into the AI contribution line.
7. Capital Structure: Financing amounts, cash received, and conversions must be distinguished.
[Fact] The announcement on September 29 was the first closing of USD 15,000,000 financing; there is also a maximum priority guarantee limit of USD 5,000,000. The conversion of related securities into preferred shares is conditional on completing the Nasdaq listing, etc.; approximately USD 6,500,000 of original notes are involved in the exchange. The amount, old debt replacement, and conversion terms cannot be directly equated with new cash received or debt being fully eliminated. [S4]
[Fact] Prior to the acquisition, ONAR's first half of 2026 had an operating cash outflow of about USD 379,000, operating loss of about USD 601,000, and interest expense of about USD 1,634,000, of which roughly USD 372,000 was non‑cash amortization of debt discount. This baseline shows that financing costs are a constraint that must be managed separately; the period also included consolidation effects from the originally acquired business. [S5]
[Inference | Arithmetic] If the seller's note of USD 7,000,000 retains the same principal for the full year, a simple annual interest rate of 8% corresponds to about USD 560,000. It represents only a portion of the funding cost, cannot stand for total interest, and the adjusted EBITDA of the target USD 6,600,000 should not be used directly to calculate debt service coverage.
[Advice] The CFO should provide a 13‑week cash forecast and a 36‑month maturity schedule, itemizing fixed debt repayments, potential earn‑out triggers, technology investments and working capital. The investment amount for new acquisitions should be determined by the cash available under a conservative scenario; potential IPO or conversion should not be treated as already realized cash relief.
8. Hundred‑Day Roadmap and Board Acceptance Form
- Day 1–15 | Retention Business: Confirm service continuity with key customers, map management personnel, customer contracts, billing rules and data usage rights. Acceptance: Key customer owners are clear and service interruptions are traceable.
- Day 16–30 | Build Baseline: Establish ledgers for revenue, service cost, collections, labor hours and quality at the customer/month granularity, and adjust contract gross profit and management contribution. Acceptance: The full profit of a set of customers can be recalculated.
- Day 31–60 | Pilot: Launch two workflows with clearly assigned business owners, recording AI inference, engineering, manual review and rework costs. Acceptance: Delivery cycle and unit cost are comparable under the same quality standards.
- Day 61–100 | Replication Test: Have a second team deploy validated components and maintain them independently. Acceptance: Additional deployment cost, customer retention and cash contribution are clear; processes that fail are held back from expansion.
- Board monthly table: customer revenue retention, gross profit per customer, revenue and contribution per effective full‑time employee, delivery hours, rework rate, cash collections, total technology cost, cumulative contract earn‑out and available cash. The denominator and statistical scope are fixed.
【Recommendation | Responsibility Allocation】CEO is responsible for expansion and capital allocation; the target’s management team handles customer and contract performance; Labs head oversees systems, reliability and handover; CFO manages cash bridging and prevents double‑counting of earnings. Shared incentives should include quality, retention and net profit to curb the impulse to chase per‑person account counts alone.
IX. Counter‑evidence and Conditions for Halting Expansion
- [Fact] ONAR regulatory materials clearly indicate going‑concern concerns, working‑capital shortfall, increased debt and integration risk. These disclosed constraints should be incorporated into the project budget and cash forecasts. [S1]
- [Fact] The target’s historical financials are unaudited and the full QoE has not been disclosed. Certain add‑backs in the adjusted profit may continue to represent economic costs. [S2]
- [Inference] If client revenue retention continues to deteriorate, financial variances cannot be reconciled, or net unit cost does not improve after two full delivery cycles, the pilot should not be expanded and the cause investigated.
- [Inference] If the second deployment still requires a comparable amount of engineering man‑days, the platform may remain dependent on bespoke projects; subsequent M&A valuations should not pre‑emptively include scale‑up technology synergies.
- [To be verified] The target’s formal financial statements, net financing proceeds and interest rate, client pipeline, on‑site Labs configuration, actual technology costs, and post‑closing cash flow. Without this evidence, the acquisition IRR or AI’s standalone contribution to profit cannot be confirmed.
[Decision conclusion] This transaction is suitable for building a long‑term contract‑to‑cash observation sample. In the short term, the most telling signals are whether the original client remains, whether the two workflows reduce full delivery cost, whether the second team can be replicated at low cost, and how much cash the group retains after paying financing and integration expenses.
Research methodology and primary sources
Verification date: October 5, 2026. All amounts are in USD. Scenarios and roadmaps represent the report’s recommendations and do not constitute company plans or guidance. The report did not conduct client interviews; press releases reproduced are not considered independent verification. FDEPE conducted an independent study and has no affiliation or sponsorship relationship with McKinsey.
[S2] ONAR, 2026-09-30, 8-K attachment 99.1; target financials, add‑backs, management team and Bloom.