Risky, but promising” – this phrase can describe any business model with a chance for great success! After all, high risk and high profitability always go hand in hand 😉 The “give-to-get” model, where you give your information to get someone else’s, falls into the same category. And here’s a real startup that increased its revenue tenfold within a year after implementing this model.
Project Essence
Compa helps companies understand what competitive salaries they need to offer candidates to avoid underspending or overpaying.
In principle, market research on labor and salary is already conducted by research companies and employers themselves.
The only problem is that they conduct research periodically—so today, companies have to rely on data from the day before yesterday. Plus, participants in the research can lie about their salaries, so this data may not be entirely realistic.
Compa’s data is free from these drawbacks—because they are collected in real-time from authoritative sources.
The mechanism for this is simple. Employer companies connect their systems, through which they communicate with candidates (ATS, Applicant Tracking System), to the Compa platform—after which:
data on offers made by the company to candidates goes to the Compa platform, in exchange, they receive data on offers made by other companies to their candidates. Naturally, data on offers from other companies is aggregated without mentioning names or other personal information about the candidates.
But why does Compa collect data on offers made, and not actual salaries? Because, as the startup claims, the offer is a “marketing interface” through which companies interact with candidates.
When a company is about to hire someone, the first thing it should do is not lose the first “marketing battle” for them—by making an offer that they won’t immediately throw away but will be willing to discuss further.
A possible problem with the salary database is that different companies may use different job titles for essentially the same position, or vice versa—use the same titles for different positions. Moreover, salaries for the same position may vary by region.
Therefore, each client company of the platform has the opportunity to set rules according to which job titles and regions from the common Compa database will be matched with positions in the specific company.
In addition, salary and offer sizes usually depend on the type of employer company. Large companies may have one salary level, fast-growing startups raising investments may have another, and small companies may have a third.
For this purpose, the client company of the platform can create several groups of companies, where they can place companies that are benchmarks for salaries—and see the averaged salaries for the required position for each of these groups separately.
The platform collects and distributes offer data among its clients in real-time. Therefore, platform clients can instantly notice and react to every change in the market situation.
The target audience of Compa is currently only technology companies. Currently, the platform contains information on more than 110,000 offers made to candidates by such well-known companies as Stripe, Nvidia, Doordash, Squarem Dropbox, Instacart, Autodesk, Vomeo, and MongoDB.
The subscription cost for a company to access the Compa offer database starts at $50,000 per year.
The startup started operating in 2021, receiving $3.9 million in investments for a corporate platform for creating offers to candidates.
The offer database is a new product that they launched in the spring of 2023. Since then, the number of startup clients has grown by 800%, and revenue has increased tenfold.
Currently, Compa has raised new investments of $10 million.
What’s interesting?
The ideological principle of Compa is the “give-to-get” business model. To receive information about others’ offers, you need to share information about your own offers.
The cost of using the offer database is formulated as “from” $50,000 per year. It is possible that this cost depends on how many offers the client makes. Thus, companies that make fewer offers should pay more money because they receive more information than they give in return.
I first noticed this model in the fall of 2021 when I wrote about the startup Varos. They made a platform for exchanging marketing and financial information between online stores and cloud services—collecting data on ad click-through rates, conversion rates to purchases, repeat purchase rates, customer retention rates, order sizes, and revenue.
At the time of the Varos review, they had just graduated from Y Combinator, receiving $125,000 in seed investments. However, in 2022, they raised $4.3 million in new investments for the development of their platform.
Two other startups, Crossbeam and Reveal, also use a very similar principle of exchanging “sensitive” information.
These startups have created platforms where B2B sellers can automatically exchange information from their CRMs with each other. The goal is to help each other establish “warm” contacts with potential customers when one company already has such a contact among its customers, and the other company just wants to reach out to them.
This model has proven to be in demand, as Reveal raised $54.3 million in investments, and Crossbeam raised $116.9 million.
Renowned venture investor David Sacks from the “PayPal Mafia” believes that the “give-to-get” model could become popular among AI startups.
By the way, it turns out that the first application of the “give-to-get” model was by the startup Jigsaw, founded in 2004. They used this model to maintain a database of company contacts. Each participant in the service earned “points” for adding, updating, and editing contact information and could spend them to obtain contacts from the database. In 2010, Salesforce acquired this startup for $142 million.
David Sacks believes that AI startups can use a similar “points” model to collect the data they need to train their AI machines—earning points for submitting their data to the central repository and spending them to obtain others’ data.
The potential demand for such a model may be explained by the fact that large sets of high-quality data are a critical component for any startup involved in AI training.
In the author’s opinion, such platforms will be very useful for data exchange among AI startups working in the fields of medicine and health, finance and investments, scientific research, manufacturing, creativity, as well as analysis and drafting of legal documents.
Where to head
A possible direction of movement is the creation of platforms for exchanging important information between companies based on the “give-to-get” model. Within this framework, two paths can be pursued.
The first path involves creating platforms for exchanging “traditional” information similar to Compa, Varos, Crossbeam, and Reveal: salaries, marketing and financial metrics, contacts of buyer companies.
Moreover, different types of companies may require different information – logistics costs, procurement prices for various categories of goods and services, and something else. By the way, what other interesting information could be exchanged using this model?
The second path involves creating platforms for data exchange in newer AI topics.
In which areas do companies currently lack data for AI training? Do other companies have such data? Is it necessary to constantly update and revise this data to maintain and improve the quality of the corresponding AI machines? How can secure data exchange be established?
The “give-to-get” model may seem risky in terms of data security. In addition, special efforts must be made to constantly ensure that this data reflects reality and is not fabricated solely for the purpose of obtaining others’ data.
On the other hand, models like Airbnb, Uber, and Blablacar also seemed very risky at one point. However, they succeeded precisely because they were risky 😉 After all, high risk and high profitability always go hand in hand.
So, one of the applications of today’s “give-to-get” model may turn out to be similarly successful. Which application of it seems most promising to you?
About the company
Compa
Website: trycompa.com
Latest round: $10M, 01/29/2024
Total investments: $13.9M, rounds: 2