SOLVE has launched cohort-based Relative Value Analysis, an extension of its pre-trade platform that lets institutional fixed income traders and portfolio managers measure a bond against a dynamically defined peer group rather than picking off comparisons one bond at a time. The release, announced this month, adds two analytical frameworks – bond versus cohort, and cohort versus cohort – to a platform the firm has built around predictive pricing for fixed income securities.
The cohort analytics sit on top of that pricing layer, which is where SOLVE locates fixed income’s core problem. “Fixed income is still very much a space defined by significant information asymmetries, persistent data fragmentation and manual process,” says Eugene Grinberg, Co-Founder and CEO of SOLVE, in conversation with Market & Alt Data Insight. “It remains a good generation or two behind equities and other markets. Everything starts with the fact that it’s very difficult to get accurate pricing on the underlying securities.”From One Bond to a Population
The cohort framework formalises a hierarchy of questions that desks already ask informally. At the base is the single-bond problem: a security is available, and the trader needs to know what it is worth. “The most basic question is: I can see a bond available in the market – what price should I be paying?” says Grinberg. “That sounds as though it should be easy to answer, but it isn’t. You may see some bids and offers, but how do you know what the correct price is? How does it stack up against other bonds that are very similar in nature, in terms of their risk profile, their duration profile?”
Above that sits the portfolio manager’s question, which is not about a bond at all but about a segment. “You can abstract further still: forget bond versus bond – is this even the right sector for me to focus on?” says Grinberg. “You set aside the bond-versus-bond analysis, look at it from 10,000 feet, and run that cohort-versus-cohort analysis instead. How does a cohort stack up to how it has historically performed against other cohorts? That’s more of a portfolio manager-level analysis, whereas the other work sits closer to the trader.” The two framings released this month map onto those two seats. One supports security selection; the other supports capital allocation across sectors and themes.Who Builds the Cohort
The peer groups are not curated indices, they are constructed by the client.. “The cohorts are fully configurable by the user,” says Grinberg. “These aren’t static indices we’re creating; we give the user full control to define a cohort.” A natural language interface lets users assemble those groupings conversationally rather than through dropdown filters.
User-defined cohorts are the source of the tool’s flexibility, letting a desk tailor a peer set to the specific question it is asking. A grouping assembled to reflect genuine duration, risk and sector comparability will surface real dislocation, and the conversational interface lowers the effort of building and refining those sets. The platform’s suggestion layer offers a starting point based on a bond’s most obvious characteristics, which users can then adjust to fit their own view of what constitutes a true comparable.
The Part That Closes the Gap
A bond can look cheap against its cohort and remain untradeable at the level shown, and fixed income is prone to that gap. “Fixed income doesn’t give you the conviction that just because a bond has a published price you can always follow through on it,” says Grinberg.
SOLVE’s answer is to attach market context to the signal, drawn from quote data the firm reports collecting for close to 15 years from messages, emails and other unstructured dealer communications. “When we present the relative value analytics that tell you a bond is rich or cheap versus its cohort, an important part of the puzzle is whether it’s actionable,” says Grinberg. “We publish metadata alongside our quotes data – the size at which it’s offered, whether it’s a one-way or two-way market, and whether it’s an axe, meaning a dealer has been specifically tasked by a client to do something with that bond. Those data points give you more conviction that this is actionable, rather than the dealer simply carrying some stale offer price on their inventory sheet.”
The metadata layer is what does the work here. An axe is a different object from a stale inventory price, and a signal that carries that distinction tells a desk more than one that does not. SOLVE reports aggregating more than 30 million daily quotes from street messages; the classification attached to those quotes – size, market direction, axe status – is what turns the volume into something a trader can act on. The consistency and accuracy of that metadata across the long tail of illiquid names is what underpins the executability claim, and a natural point of focus as the capability is used in practice.
The wiring also places the launch within a visible competitive moment. ICE this month launched ICE Compass, an AI-driven pre-trade platform that estimates where a fixed income trade is likely to clear and ranks dealers on pricing competitiveness, addressing the information leakage that comes from revealing intent in a bilateral market. SOLVE and ICE are solving different parts of the same pre-trade sequence – Compass is concerned with where and with whom a trade clears, SOLVE with whether an identified opportunity is real and actionable in the first place – but the common thread is unmistakable. Confidence at the point of decision, not raw analytics, is where fixed income pre-trade tooling is now competing.
Some Firms Need the Data, Some Need the Answer
The launch also says something about who it is for, and here Grinberg is direct about the market’s stratification. “Some of the leading firms have had data and AI as part of their culture for years,” he says. “They’re really pushing the ceiling on what performance could look like in fixed income, with more active portfolio management rooted in data and automation. That’s creating a lot of pressure on the long tail of firms that haven’t driven this kind of change.”
He continues: “The leading firms may not need our relative value analytics – they buy our data and transform it into trading ideas with their near unlimited resources. But most firms don’t have unlimited resources. One of our internal mantras is that some firms need the data and some firms need the answer. The relative value analytics are the answer. The leading firms need the data. Everyone else needs both.”
SOLVE is positioning cohort-based relative value as infrastructure for the first group and as a productivity layer for the second. The analytics are the visible product. The data supply chain underneath them – the quote history, the metadata, the predictive pricing the cohorts are measured against – is what will determine whether the answer the platform hands its clients is one they can act on with confidence.
Subscribe to our newsletter
