The Dynamic Landscape of Venture CapitalThe dynamic landscape of venture capital hinges significantly on the effective alignment of potential investment projects with a fund’s defined thesis profile.Among a multitude of factors that constitute this alignment, discerning the industry sector of prospective investment opportunities is particularly challenging, yet unequivocally crucial.Traditionally, this task has been entrusted to skilled human analysts, a process that involves deep-diving into each investment opportunity to extract pertinent information (Kaplan and Stromberg, 2001). While human analysis has been lauded for its accuracy and in-depth understanding, it is also associated with considerable time investment, financial expenditure, and inherent limitations when it comes to scalability, which has led to calls for a more efficient, automated methodology (Hochberg, 2016).
Read Full Research
We aim to refine a multi-label classification methodology to achieve more accurate industry/sector categorization based on textual project descriptions.In the context of multi-label classification, it is feasible for a single project description to correspond to multiple industry sectors. This introduces intricate challenges, not only in terms of adapting the language model but also in identifying suitable metrics to evaluate its efficacy.
Evaluation of Untuned GPT-3 Turbo 0301In this report, we offer a comprehensive evaluation of the capabilities of untuned GPT-3, focusing particularly on the GPT-3 Turbo 0301 variant.
Read Full Research
Understanding the dynamics and relationships within the startup ecosystem is crucial for entrepreneurs, investors, policymakers, and researchers.
In this research, we employ unsupervised learning techniques, specifically clustering and dimensionality reduction, to classify startup projects into meaningful domains. We assemble a comprehensive dataset of startup information from various sources, including Crunchbase, and fine-tune a T5 model to classify industry sectors based on text descriptions. With a dataset of approximately 100,000 observations, we explore dimension- ality reduction techniques to identify the most informative features for clustering. Using the K-means algorithm, we determine the optimal num- ber of clusters and apply it to the startup dataset. Our results reveal four distinct clusters, demonstrating meaningful separations among star- tups based on their attributes. The t-SNE visualization technique aids in understanding the relationships and patterns within each cluster. This research contributes to a deeper understanding of the startup ecosystem, empowering stakeholders to make informed decisions and foster the growth and success of startups.
Read Full Research
Background
AutoKitteh, a stealth tech company, had caught the attention of numerous venture capital firms. Unfortunately, many potential investors, including Cervin Ventures, were left in the dark until after the investment round was closed.
Challenge
Our challenge was to retroactively determine the depth and timing of interest from various venture capital firms in AutoKitteh using our advanced analytics software.
Results
Our software successfully pieced together a timeline of expressed interest from various stakeholders:
Insight Partners, April 2023
Work-Bench, April 2023
F2 Venture Capital, April 2023
Index Ventures, April 2023 (reiterated interest in May 2023)
Cerca Partners, May 2023
Conclusion
Beyond just the timeline, our software was able to gauge the intensity of the interest, determined by factors such as the frequency of signals and the prominence of activities. For instance, Index Ventures showcased a particularly high level of engagement, backed by their two separate expressions of interest in April and May 2023.
Background
Cervin Ventures put our advanced analytics software to the test. They challenged us to uncover a snapshot of their recent investment pipeline, scrutinizing the relevance and timeliness of our software's signals.
Approach
We deployed our software, designed to analyze an extensive array of data sources from web analytics and social media interactions to public filings and direct communications. The software uses machine learning algorithms to recognize patterns and extract signals indicative of potential investment interest.
Results
In a live demo call, our software unveiled signals related to a range of companies that Cervin Ventures had recently shown interest in, achieving an impressive 100% accuracy. The list of companies includes:
Altostra
Cobalt
Elisity
Opaque
Bleach
InKeep
Radiant Security
What made these signals truly valuable was their extreme relevance and timeliness. They represented recent activities, offering a window into the investment pipeline in real-time. If the demo call had been scheduled earlier, the signals would have been even fresher, providing Cervin Ventures with a potential lead time advantage.
Conclusion
Our software's unique ability to uncover and interpret relevant, timely signals offers a competitive edge in the high-stakes world of venture capital. As demonstrated in the challenge set by Cervin, our tool provides actionable insights into recent investment activities, enabling venture capitalists to make informed decisions swiftly and efficiently
Background
Comfort Del Gro Capital Partners, having recently made a strategic investment of $4.3m into the fund Shift4Good, presented us with a distinctive challenge. The fund aimed to gain access to deal flow, and our task was to utilize our software to demonstrate that objective's realization.
Approach
We put to work our advanced machine learning capabilities, which sift through a sea of data from various sources, to decipher patterns and detect signals indicative of potential investments.
Results
In a live demonstration, our software unveiled Shift4Good's deal-flow. The signals were immediately validated by an analyst from CDGs, attesting to the 100% accuracy of our tool. The signals pointed to the following companies:
Clearly
AnyRoutes
Addionics
Tolv Systems
Lyko Global
Nelson Mobility
Conclusion
Instead of investing millions of dollars for the same information, our advanced analytics can provide a more cost-effective and efficient solution, delivering invaluable insights directly to venture capitalists' fingertips.
Background
Mobility Fund, one of our early clients, had made two significant investments since joining our platform in Q4 2022. The challenge was to utilize our software's predictive capabilities to unveil the intent behind their investments in Pionix and Colonia.
Results
In January 2023, our software detected early signals of Mobility Fund's interest in Colonia from one of the fund's senior members. To strengthen the signal interpretation, our software algorithm cross-referenced the fund's investment thesis and discovered overlapping interests from other prominent European investment funds:
Atlantic Labs
AutoTech Ventures
UVC Partners
Freigeist
Significantly, our software uncovered a robust relationship between FlixBus, a previous investor in Colonia, and Mobility Fund. This additional insight solidified our prediction of an imminent investment from Mobility Fund in Colonia. Our predictions were validated when Mobility Fund participated in an investment round that also included:
Vent.io
Plug and Play
Atlantic Labs
Octopus Ventures
Grover
Conclusion
This case highlights our software's unique ability to discern key deal sources and sinks, providing a reliable prediction of a venture fund's investment intentions. It demonstrates how our technology can unveil direct activity signals while also extracting valuable insights from the broader investment community.
Our clients include private equity, venture firms, and corporate venture arms.
No, TwoTensor focuses solely on signalling live deals.
We rigorously validate our signals monthly against newly announced deals, offering real-time access to our clients. Our case studies with various venture funds showcase live validation, and we encourage skeptical prospects to experience our technology firsthand through test samples.
Yes, our platform scales globally through tech-driven data collection and strategic partnerships, ensuring insights in even newly entered markets.
We automate deal sourcing by identifying investor-active deals and matching them to your investment criteria using AI, streamlining the process end-to-end.
Our pricing is tailored to match the specific needs of customers, including data requirements and active sourcing mandates.
We rigorously monitor aggregated and anonymized public data. We do not deal individual or private data. Our value lies in inference.
We process real-time data to predict future events rather than cataloging historical events.
Yes, our platform supports seamless integration with major CRM systems, including Salesforce, through API access for efficient workflow management.
Qualified clients can access trial data or sample reports, allowing them to evaluate our insights' depth and quality. To protect the interests of our paying customers we do not offer free access to live data.
Our platform provides global deal flow, which users can refine based on their investment thesis for targeted deal insights.
We invest in technology and expertise to ensure our platform's scalability and data quality, with customization options to meet diverse client needs across all sectors and regions.
We automate deal sourcing by identifying investor-active deals and matching them to your investment criteria using AI, streamlining the process end-to-end.
We launched our operations in Q4 2022, bringing a fresh perspective to venture capital analytics.
Our deal signals are updated monthly, reflecting the dynamic nature of our analysis and the venture capital landscape.
The past five years have seen a dramatic transformation in data availability. With advancements in AI, we now have the ability to conduct contextual analysis on a vast scale, making it an opportune moment for leveraging alternative data in venture capital.
This primary brochure provides an overview of our business and operations.
Main Brochure PDF
This brochure provides an overview of our performance in 2024, with a focus on the fourth quarter.
2024 Backtesting
This one-pager provides an in-depth review of intent data, the primary focus of TwoTensor.
TwoTensor One Pagers
Eagle Alpha is a premier platform for alternative data, offering over 1,850 data products to enhance alpha generation for investors and firms.
https://www.eaglealpha.com/
Neudata.io offers alternative financial data discovery for investment firms, enhancing market strategies.
https://neudata.co/