Yardstick Research tear-sheet / AI sales cohort

Methodology · how we score · rubric weights in plain sight · vendors received this sheet seven days before publication and could flag factual errors, never rankings

Tofu

Identity

• Founded: [UNKNOWN - not disclosed on vendor site]
• HQ: [UNKNOWN - not disclosed on vendor site]
• Legal entity: Tofu Technologies, Inc. [VENDOR-CLAIMED - tofuhq.com]
• Funding: [UNKNOWN - not disclosed on vendor site]
• Headcount: [UNKNOWN]
• Recent news (last 12 months): No 2025 news items disclosed on vendor site as of 2026-05-27 [VENDOR-CLAIMED - tofuhq.com]
• Archetype: AI content personalization layer - sits on top of existing martech stacks (Salesforce, HubSpot, Marketo, Outreach) to automate always-on, multi-channel personalized campaigns without rip-and-replace

Total score: 31.7 / 100

Scoring: equal-weight mean of 6 dimensions × 100, less pricing-transparency penalty (hard = 10 pts).

• Stage fit:
• Foundation (<$5M ARR): no - Tofu's value proposition assumes an existing martech stack with Salesforce, HubSpot, or Marketo already deployed; Foundation-stage companies without that infrastructure will not benefit from a personalization layer on top of it
• Pilot ($5-$50M): yes - SOC 2 and enterprise-targeted positioning notwithstanding, Pilot-stage SaaS companies with HubSpot + Outreach already running can layer Tofu on top without replacing their engagement stack
• Scale ($50-$500M): yes - the primary fit; Scale-stage teams running multi-channel campaigns across Salesforce, HubSpot, Marketo, and Outreach are the stated ICP; the personalization layer model maps well to companies that have sequence infrastructure but need AI-generated content variation at volume
• Optimization ($500M+): conditional - SOC credentials signal enterprise-readiness, but no named enterprise customer proof, no published pricing, and no analyst recognition as of this writing create procurement friction at Optimization-stage buying processes
• One-line verdict: A personalization-layer play for Scale-stage martech stacks; the non-rip-and-replace positioning is differentiated, but zero pricing transparency and reviewer-flagged analytics gaps mean buyers must ask hard questions before signing.

Headline numbers

Dimension scores

Equal-weight scoring - 6 dimensions, each 16.7% of the base score.

Pricing detail

Tofu does not publish pricing. There is no pricing page in the site navigation as of 2026-05-27, and all primary CTAs route to "See Product Demo." No historical pricing disclosure is available in the Wayback Machine. Buyers should assume sales-led pricing with enterprise-tier TCO given the SOC-credentials and Salesforce/Marketo integration positioning. [VENDOR-CLAIMED - tofuhq.com; MEASURED - D1, yardstick-data pricing_transparency_note]

Integrations

From D1 catalog [MEASURED - D1, yardstick-data]: - Salesforce Sales Cloud Einstein - Salesforce Einstein Copilot - HubSpot Breeze Copilot - Outreach

From vendor site [VENDOR-CLAIMED - tofuhq.com]: - Marketo

Tofu is explicitly positioned as a personalization layer that works across existing martech rather than replacing it. The four-stack (Salesforce, HubSpot, Marketo, Outreach) integration footprint is well-suited to mid-market and Scale-stage B2B companies. [VENDOR-CLAIMED - tofuhq.com]

Editorial assessment

Tofu positions as a non-rip-and-replace personalization layer in the content and personalization category. Rather than competing with the engagement platforms (Outreach, Salesloft), Tofu sits one layer above them - ingesting CRM and intent data, generating personalized content variations, and pushing campaigns through whatever engagement stack the buyer already runs. The wedge is reasonable: enterprise buyers with deployed Salesforce + HubSpot + Marketo stacks resist platform replacement but will add a layer that promises incremental lift. The SOC 2 certification indicates the company has been through enterprise security review and has customers who required it. [VENDOR-CLAIMED - tofuhq.com; ESTIMATED - Yardstick Research scoring rubric]

The main weakness is opacity. No pricing is disclosed, no customers are named on the public site, and no G2 or analyst recognition is verifiable as of this writing. The deliverability infrastructure score of 2/4 reflects that Tofu operates as a content layer and delegates deliverability to the underlying engagement platform; the architecture is reasonable, but deliverability risk lives in the platforms Tofu does not control. The setup time score of 1/4 reflects that integrating a personalization layer across Salesforce, HubSpot, Marketo, and Outreach simultaneously requires meaningful integration work and a RevOps owner to own the configuration. [ESTIMATED - Yardstick Research scoring rubric]

Reviewer feedback on record (from D1 top_gap field) flags missing analytics: Tofu-generated asset performance is not natively trackable within Tofu - buyers must instrument attribution through the underlying engagement platform. For Scale-stage RevOps teams that need to justify personalization-layer ROI to leadership, this creates a measurement gap to surface in pre-sales diligence. [MEASURED - D1, yardstick-data top_gap field]

Best for

• Stage: Pilot ($5-$50M ARR) and Scale ($50-$500M ARR)
• Company profile: B2B SaaS companies with an existing four-stack (Salesforce or HubSpot + Marketo + Outreach) that want AI-generated content personalization across all channels without replacing their engagement platform; RevOps teams that have sequence infrastructure in place but spend manual time creating content variations per-persona or per-account; companies with procurement teams that require SOC 2 before approving new martech
• Skip if: The company does not yet run Salesforce, HubSpot, or Marketo - Tofu has nothing to layer on top of; buyers who need published pricing for budget approval; teams that need native attribution and campaign-analytics within the personalization tool itself; Foundation-stage companies without an established engagement stack

Right-of-reply

Tofu received this tear-sheet seven calendar days before publication of the Yardstick Research 2026 Yardstick Report, including all measured numbers, sample outputs, and editorial assessment. Tofu was given the opportunity to flag factual errors - incorrect pricing, misquoted feature availability, outdated screenshots, factual misstatement in the editorial assessment. Tofu was not given the opportunity to request a score revision, dispute the rubric or its weights, withdraw from inclusion, negotiate ranking placement, or suggest changes to the editorial assessment beyond factual correction. Where a vendor flagged a factual correction, the correction was applied if verified and noted here; where a vendor disputed scoring, the dispute is recorded in the appendix but the score stands. Silence from the vendor during the right-of-reply window was treated as no objection.

Sources

• https://www.tofuhq.com/ [VENDOR-CLAIMED - homepage, accessed 2026-05-27]
• D1 database: yardstick-data, vendors + vendor_dimension_scores tables [MEASURED - D1, yardstick-data]