Thematic Research Compilations
- Problem Solved: Sales and Research can only compile thematic summaries of content manually and if they know everything published in detail.
- Client Benefits: Automated compilation of documents according to themes.
- Ingestion & Atomisation
- Post-Publication
- Automatic Tagging
- Document Level Tags
- Output
- Client Type
- Sales
- Where does the Client use it?
- Smart Marketing
Compilation Emails are the unsung heroes of Research Distribution – we can make them even better!
Every Sell Side firm has a product (or many versions of a product) which, in its most basic version, is just an email with a long list of hyperlinks from which a recipient can click through to any of the day’s published research reports. This might be called a ‘Morning Email’, a ‘Morning Espresso’, a ‘Daily Research Highlights’ or any multitude of names, but it always performs a similar function.
The sheer volume of reports published by Sell Side Research departments every day dictates the need for this compilation email. Sales desks rely heavily on it so that they always know what Research has been published and can then repurpose the email for their own needs, moulding it into customised insights which can go to their clients.
The production value varies across firms. Some look like glossy marketing emails, while others are given a deliberately rough-and-ready feel and might be updated throughout the day. The method of compilation also varies. Some are produced with limited human interaction and use simple automation to choose the content and send them out. Others require a small, dedicated team pulling it together manually every morning before 6am, the human control allowing for total subjectivity when it comes to choosing particular documents to highlight or other information to add.
But what do they all have in common? The fact that they could do so much more if they had access to better tag data.
At the moment, the organisation of the Research and other information in these products is done according to the traditional tagging: Sectors, Analyst Names, and Regions tend to provide the main structure. But the true value of Research emerges when it cuts through those verticals. Themes or Topics or cross-Asset ideas are where the Research experts can make a real difference to their clients’ investment decisions when speaking to them 1×1, but there is currently no way for the written Research itself to be consumed that way.
Enter Limeglass.
Having tagged each document with its massive lexicon of financial terms, Limeglass can provide you with >10x more metadata on each document, allowing for those documents to be treated in any number of previously impossible ways.
Starting with the simplest first, just adding selected tags to a compilation product is one way that some Sell Side firms have gone to give Salespeople and Clients more useful information about which documents they should read.
In the example document second from the top, there is nothing in the traditional metadata for the document that would suggest that a reader interested in effects of a loosening labour market might have on equity valuations. But Limeglass is able to identify the #Labour Market topic being written about in the body of the document and then expose it as a tag.
In a more advanced version of this use case, it would be possible to make that tag clickable, so that a reader can use it to find more research on that theme. Taking another example, the third document about CompanyOne has a tag for #Disruptive Technology. Clicking on that tag could take the reader to a different view / section of the email which focussed on that theme:
Suddenly, the structure of how information is displayed on the page can be changed from the original ‘Highlights’ and ‘Sectors’ sections, and replaced with a focus on the Disruptive Technology theme, split between today’s research and older documents. That is just one example, of course.
Not only does this give you a whole different lens through which to view this report on CompanyOne, it could also be set up to show the specific paragraph(s) inside the document related to the theme. In the example above, a reference to ‘Artificial Intelligence’ is what flagged the document. In the next document down, it was ‘Internet of Things’.
Identification of Topic Correlations at paragraph level
- Problem Solved: Difficulties in finding tradeable instruments thematically.
- Client Benefits: Automated detection of co-location of specific topics in individual paragraphs allows for the discovery of cross-asset thematic links.
- Post-Publication: xx
- Paragraph Level Tags: Limeglass atomises your documents, breaking them down into individual paragraphs, tagging each one in great detail meaning co-location of tags can be tracked while preserving the context in the original text.
- Ingestion & Atomisation
- Post-Publication
- Automatic Tagging
- Paragraph Level Tags
- Output
- API
- Client Type
- Sales
- Where does the Client use it?
- Data Science
Specific Client Use Cases:
- Sell Side Analyst collaboration using Key Research Indicator tags
- Set up alerts for certain key tags so that analysts can know when their colleagues are writing about things that might have a material impact on their own coverage
- Buy Side Tradeable Instrument ideas
- Track the co-location of certain topic tags with tradeable instrument tags. For example, any stocks mentioned in the context of an important theme.
Helping you connect the dots.
Structure your Unstructured Data!
Investment Research is, by its nature, a curious combination of highly structured information (Macro or Micro financial forecasts, Equity Ratings, financial ratios) with completely unstructured information (investment theses, background information, company descriptions).
Some of the structured information can be used to ensure that people are reading the right content and that certain important connections have been picked up. When an Oil analyst publishes a piece updating their Brent Crude forecast, for example, if that forecast is stored in a database, it can be flagged to colleagues (perhaps Airline analysts and Economists) who rely on it for inputs to their own models.
However, there are plenty of situations where that sort of useful automated cascading of information is impossible because the information is not structured. And this is where paragraph-level topic correlations comes in.
When an Economist writes about the relationship between consumer spending and jobless claims, you would hope that the Economist’s Consumer Stock analyst colleagues would read the research to understand hard or soft impacts on their forecasts. Do they need to update the growth rates plugged into their DCF valuation models?
Sell Side Analysts are Resource-Rich but Time-Poor
But even though they theoretically have this information available to them, they are unlikely to read it. Analysts are resource-rich but time-poor. And, given that the information they need is unstructured text, there is no way to set up a system to flag it to the right people.
However, if that document has been processed and atomised by Limeglass, the content tags for each paragraph will be readily available and an automated system could be built off that information. ‘Consumer Spending’ and ‘Jobless Claims’ would be separate tags that would both occur in this paragraph.
If the Consumer Stock analysts were set up to receive alerts for the ‘Consumer Spending’ tag, they would be made aware that there is a potentially interesting co-occurrence with the ‘Jobless Claims’ tag. This may pique their interest enough to read the relevant paragraph, thus improving the information flow between different subject matter experts on the Sell Side.
Buy Side clients want to see “joined-up thinking” from their Research providers
The same logic works for the actual consumers of the research as well: the Buy Side. A fund manager will be much more likely to take research seriously if they can see that their Sell Side counterpart is making collaborative investment recommendations with joined-up thinking. If they are reading the Economist’s report and can be alerted to ‘related research’, they could click through to the Consumer analyst’s report which references the other’s work.
A system that can achieve this is not any ordinary NLP tagger. Limeglass can do this because of its use of two key proprietary technologies: Atomisation and Rich NLP.
Atomisation breaks documents into individual paragraphs and then preserves all the relevant metadata and context whenever these paragraphs are then used. Rich NLP, unlike most NLP tools, uses knowledge graphs based on both the plain text extracted from the document, and important context like positioning on the page, formatting, and the use of headings and sub-headings. It is only by bringing these two systems together that you can analyse tag co-location in a meaningful way.
Automated Pre-Publication Document Tagging for Enhanced Research Distribution
- Problem Solved: Inability to tag the main Themes of a document at scale. Only current solutions require error-prone manual tagging.
- Client Benefits: Ability to distribute research in Thematic categories and to create tailored thematic content.
- Pre-Publication: Limeglass Prism automatically tags your documents before they have been published, ensuring seamless progress to distribution.
- Document Level Tags: Surface relevant tags from the content of your document and give your analyst the power to select or deselect them.
- Ingestion & Atomisation
- Pre-Publication
- Automatic Tagging
- Document Level Tags
- Output
- Heatmap
- Client Type
- Research Operations
- Where does the Client use it?
- Distribution System
The New Gold Standard in Research Distribution.
Your distribution system needs upgrading
You could produce the best research in the world and yet it could go entirely unread without a first-class distribution system. Your clients on the Buy Side may occasionally ask you directly if you have written about something they want to investigate, but, in most cases, they will only read what they have seen enter their inbox.
Clients are still drowning in a flood of research reports
Most Sell Side Research operations now have a system that automatically tags and distributes documents but only does so with a fairly modest level of sophistication. Clients can choose whether or not to receive a particular analyst’s research, or they can consume content from a selection of sectors, regions, and asset classes. But they are still experiencing a high degree of information overload and are currently unable to filter their preferences using meaningful criteria based on the actual content of the reports.
As an industry, we all need to find ways to help clients reduce the volume of received reports further, without the risk of them missing out on valuable insights.
Thematic distribution solves the problem
So, what is a first-class distribution system? One that allows clients not only to filter their consumption preferences by analyst name, sector, or region, but also by an intelligent understanding of the contents of the reports: by subject matter and themes.
We have written before [insert link to Ontology article] about the ESG-focussed client who must currently subscribe to all reports from the Automobile sector when, in reality, all they need is the subset of reports about Electric Vehicles. The same could be true for an FX trader who is interested in how major European corporations are hedging their Dollar exposure, but otherwise has no need for stock reports on Bayer or Airbus.
Smart Tag It!
The distribution system required to make this possible is one that includes an intelligent auto-tagging capability during the publishing process.
Limeglass PrismAPI does exactly this. Using Rich NLP and a large, ever-growing taxonomy of financial terms, Analysts can finish writing their document, enable the tagging system, and receive back suggested tags based on the content of the document.
Publishing Analysts can still control their document tags
In our earlier example, the analyst who covers Airbus will be given the option to include a “EUR/USD” tag in the official metadata of their document which indicates that they have written some content about euro/dollar. Naturally, they will also be given other more obvious tags related to the company, like “European Aerospace”.
Many clients will already be subscribed to research of “European Aerospace” and will continue to receive this report through that channel. But the FX trader client can now subscribe to “EUR/USD” and receive this report too.
The inference of this across all sectors, asset classes, and types of research is huge. Suddenly, clients can use a wide array of previously non-existent distribution channels to receive only the research that they need.
Limeglass Prism ensures compliant handling of pre-published information
Of course, given that this process happens to the research report before it is officially published and disseminated means that the report is still Material Non Public Information (MNPI) at this stage. The data, therefore, cannot be stored or be accessible by anyone other than the analyst and relevant compliance / Research management teams.
Limeglass ensures this by keeping the data encrypted in transit and everything performed in-process and in-memory, leaving nothing stored in persistent media, and keeping documents inaccessible to Limeglass staff (or anyone else).
And Thematic Distribution is just the start…
Of course, once you have this pre-publication thematic tagging system set up, the sky is the limit for further use cases.
Go to the Use Case menu for an overview or click Next to see what else your new tagging system will give you.
Post-publication document tagging to validate and extend the Research metadata in your database
- Problem Solved: Inconsistent manual tagging leading to incorrect tags stored in database.
- Client Benefits: Maximise thematic discoverability of research.
- Post-Publication: Simply add Limeglass as a regular recipient of your research and get back tags for all your documents.
- Document Level Tags: Limeglass maps to your existing taxonomy or can provide an even richer level of tagging according to its huge proprietary ontology. And even if you are only set up for Document Level (rather than Paragraph Level) tags, you will still benefit from the Atomisation technology due to the Coverage metrics Limeglass calculates from the tags in each paragraph.
- Ingestion & Atomisation
- Post-Publication
- Automatic Tagging
- Document Level Tags
- Output
- Custom Delivery
- Client Type
- Research Operations
- Where does the Client use it?
- Research Portal
Transform your portal’s discovery and search capabilities with Limeglass tags.
The easiest way to benefit from thematic tags is to send Limeglass your published research
Automatically capturing intelligence about the content of a document is highly valuable both before and after the point of publication. Doing this pre-publication is arguably the best way of doing this because it gives you the ability to enhance the direct distribution of your research to clients.
But if you are looking for the benefits of automated tagging without integrating it into your distribution system, the Limeglass post-publication option offers you countless options for this. And the great news is, you can get set up simply by sending your already-distributed research to Limeglass and start receiving tag data back.
One use for this that has been popular with Sell Side firms is to enhance the discoverability of research content on their portals. Existing treatment of content on portals tends to be done in two ways:
- Automation using simple rules based on existing metadata. It is fairly easy to use whatever simple tagging metadata is available to display reports in simple categories. For example, an Industrials Equity Sector page can easily be populated with Industrials research reports.
- Manual curation. This is often done to highlight the highest quality pieces of research or perhaps to group documents together under a single theme. At the moment, it has to be done manually because there is insufficient metadata to automate it.
Don’t manage your portal pages manually – use Tags
Let’s take ‘Artificial Intelligence’ as a theme to which the Sell Side might want to devote a portal page. A glance at Limeglass’ results on a random day for documents mentioning AI quickly reveals results from Consumer, Banking, Industrials, Software, Hardware and many more. It even touches documents from multiple asset classes outside equities. Using the Limeglass ‘AI/ML’ tag as the relevant piece of metadata, therefore, allows those documents to be automatically displayed on that portal page. It is unlikely that the manual curation approach would find all of these results and display them. And if Salespeople or Clients use this portal page, they are likely to be missing key insights into AI as an investment theme.
Tags also make your Research more searchable
The other growing usage of post-publication document tagging is Search. At the moment, searching for Research documents on Sell Side portals has several drawbacks:
- The actual contents being searched through tends to be smaller than the full contents of the document. Therefore, if you search for ‘Board Diversity’, you will miss a document that uses this term on page 2, because only the contents of the first page is accessible to the engine. Or it might be even more limited, perhaps making only the title and abstract accessible, and meaning even a document with ‘Board Diversity’ mentioned on page 1 is missed.
- In the rare cases that the Sell Side firm makes the full document text accessible, searching for ‘Board Diversity’ will still miss all the documents which mention ‘composition of the board’ and ‘women in leadership’. This is because most search systems use simple text matching. This requires the user to think of every possible search term and type them in one by one.
Limeglass tags, however, provide a wonderful solution for this, again with minimal integration cost:
- It does not matter if it is difficult to make the full contents of the document accessible. Limeglass has already tagged all the documents, meaning that a user will know they will get back results for all documents using the term ‘Board Diversity’ anywhere in the content.
- The user does not have to think of all the possible search terms and go through them consecutively because Limeglass has a sophisticated tagging system that does not rely on simple text matching. Instead, the system is built on a carefully managed ontology of over 150,000 tags and ___ associated phrases. This means that the user inherently searches through the tags, rather than through the documents themselves.
Personalised Research Suggestions
- Problem Solved: Research suggestions are too wide or too limited based on available document tags.
- Client Benefits: Enriched document tagging allows more granular categorisation of research and therefore of suggestions.
- Post-Publication: Simply add Limeglass as a regular recipient of your research and get access to Limeglass’s enormous financial ontology of tags, allowing you to make your research suggestions more relevant to clients.
- Document Level Tags: Limeglass expands the number of highly relevant tags associated with your research documents, allowing useful suggestions for documents with similar tags.
- Ingestion & Atomisation
- Post-Publication
- Automatic Tagging
- Document Level Tags
- Output
- API
- Client Type
- Advisory Operations
- Where does the Client use it?
- Smart Marketing
Give clients relevant results when sending them suggestions for research you think they should read.
Limeglass can give you the suggestive power of Spotify!
Consumer platforms like Amazon, Spotify, and Netflix have, for a long time, been very good at making relevant suggestions to their users for new content to consume or new goods to buy, based on their previous usage and consumption habits. Sell Side Research providers have not yet been able to replicate the quality of these suggestions.
When confronted with a hyperlink that says “Based on your interests, you may want to read…”, most consumers of Sell Side Research know that they are likely to be shown other reports on the same stock, same currency, same asset class, or perhaps just to other reports written by the same author. In certain cases, this might be enough, but it does not unlock the same cross-selling power that Spotify achieves when it uses all sorts of data to suggest a new song.
Limeglass is here to change that for the Sell Side.
Using its gigantic financial ontology (which now comprises over 150,000 tags), Limeglass dramatically expands the number of tags attributed to any given document. In comparison to the tags normally generated by a Sell Side publisher, Limeglass typically returns at least 10x as many (and sometimes that could even be more than 50x) by analysing the contents of the document.
You need better data to make useful suggestions
The expanded set of Limeglass-generated tags includes things that the Sell Side is mostly unable to capture, such as themes.
An analyst’s Japan Economics report that talks about the effects of Industrial Automation on the labour force will be able, under existing Sell Side systems, to suggest that people interested in that report should read more by the same analyst, or perhaps more about Japanese Economics (these are fairly easy tags to capture simply from document templates).
But it is only with Limeglass that an interested reader might also be prompted to click through to more content about ‘Automation’ or the ‘Labour Market’ as themes (tags which can only be captured through an understanding of the contents of the document).
Furthermore, it is not just the massive expansion of available tags that Limeglass provides: it is also the ‘Coverage’ metrics that accompany these tags.
While most tagging systems allow only for binary tags, Limeglass provides the Coverage to show how much of a document is relevant to the tags it finds. A single reference to ‘Quantitative Tightening’ in the aforementioned Japan Economics document, therefore, would be given only a low Coverage score, meaning that it could easily be ignored. The ‘Automation’, ‘Labour Market’, ‘Japan’, and ‘Economics’ tags, however, would all get high coverage scores, and would therefore not be ignored. The suggestion system could therefore prioritise highlighting ‘Automation’ related suggestions over ‘Quantitative Tightening’ suggestions (or indeed could simply remove the latter).
Discovery of trending themes
- Problem Solved: Difficult to identify trends in recent Research topics.
- Client Benefits: Limeglass heatmaps provide an easy visual representation of key topics and can be used as a filtering mechanism to identify research documents on those topics.
- Post-Publication: Once your research has been published, Limeglass can process your documents and calculate coverage metrics for all the topics you have written about.
- Paragraph Level Tags: Tagging your documents at paragraph-level allows Limeglass to expose themes that might be mentioned incidentally across multiple documents.
- Ingestion & Atomisation
- Post-Publication
- Automatic Tagging
- Paragraph Level Tags
- Output
- Heatmap
- Client Type
- Sales
- Where does the Client use it?
- Smart Marketing
Heatmaps show you and your clients the most important themes currently affecting the investment landscape.
Themes evolve rapidly – you need a tool to keep up!
It can be hard for Research Managers to coordinate the thematic display of information disseminated to clients. When Central Bank action or a major geopolitical event takes over the narrative, it often affects all asset classes and regions (in one way or another). Research Managers would always like to respond to these events with timely, perfectly coherent strategies for their clients, with bundles of relevant research ready for them to consume.
The manual approach to thematic compilation is unsustainable…
In reality, this only happens in very special circumstances when Herculean levels of manual effort are thrown at the problem. Generally, things move too quickly and dissonance across asset classes, regions, and sectors is too great to coordinate this proactively.
…but the client thirst for thematic compilation is unquenchable.
According to Bloomberg, Thematic ETF assets could be larger than GICS sector ETF assets by 2026. Themes matter. Especially when they surprise the market and everyone needs as much information as they can consume as quickly as possible.
Clearly, the manual approach will be unable to achieve this, but, if your research is being tagged at paragraph-level by a system that can identify over 150,000 unique topics, you might not need to do this manually.
A picture tells a thousand themes.
The Limeglass Heatmap is a visually appealing tool which shows how much any single topic is being written about in your research. If everyone in a research department (from the Chief Economist to the Junior Equity Analyst covering small-cap consumer stocks) starts writing about Chinese Inventories, this will show up as a large, dedicated polygon in the Limeglass Heatmap. If prepared this way, clients can then click on that polygon to read through all the relevant research.
Alternatively, another way at looking at what is trending is to surface unusually hot themes. Central Banks and Chinese Inventories might be generally well-covered topics, and perhaps some clients would want to see what topics are ‘spiking’ above their usual coverage levels.
A specific country’s economic recovery, for example, is (hopefully) unlikely to be a long-run recurring theme, and even in a time of heightened relevance, its coverage may get dwarfed by long-run themes such as the Fed. But using an ‘Emerging Trends’ heatmap will adjust for a baseline of recurrence, and will reveal spikes in ‘Sri Lanka Economic Recovery’. Again, this will show clients all research mentioning this topic across all asset classes and regions.
