London-based Feedstock is operating a research platform that offers buy-side fund managers a one-stop-shop to managing their research activities under MiFID II.
Feedstock acts as a productivity tool by integrating with a buy-side firm’s email system. By filtering through the hundreds of research reports that are sent on a daily basis, the system prioritises only the most relevant research.
As everything is powered by machine learning behavioural algorithms, it means that over time, the more research a fund manager consumes, the more detailed a picture of their research preferences begins to emerge.
“Under MiFID II, firms will need to devise an internal system to assess the value of a sell-side institution’s research and attribute that value to a dollar amount that we believe will have to be set at the start of the year, not retrospectively,” says Feedstock co-founder Lucas Wurfbain (pictured).
“With Feedstock, we’ve created a tool that allows managers to categorise and filter information in a more coherent and actionable manner. Where the MiFID II element comes in is that the platform allows the user to generate an audit trail without having to build a compliance system that interferes with existing workflows.”
Since Hedgeweek last spoke to Feedstock a little over a year ago, the beads of sweat on fund managers’ foreheads have steadily grown as they try to prepare for life under MiFID II. During that period, Feedstock has worked on the back-end of the platform to build out its artificial intelligence capabilities, hire new talent, attract new clients and grow the capital base.
“We have just finalised a venture capital funding round with Illuminate Financial Management (IFM). Their in-depth knowledge of the capital markets fintech sector combined with their existing strategic partnerships make them the ideal investor for Feedstock,” says managing director and Feedstock’s other co-founder, Charlie Henderson.
“We’ve been working incredibly hard to develop something that we believe is revolutionary in terms of information management. To illustrate how the system has evolved over the last 12 months, we now have artificial intelligence systems that automatically recognise different types of information and prioritise it for individuals, which is quite unique, especially in the financial services industry,” he adds.
MiFID II has started to crystalise such that fund managers only have 10 months to prepare for it. Wurfbain says there is still a lot of head scratching going on as to how to implement MiFID II regulations as they pertain to investment research and the payment of that research.
“One thing that is pretty universal across all the client conversations we are having is that a whole new set of data is needed to adhere to MiFID II.
“There is still some guidance that needs to come out on exactly how the buy-side should pay for research but what is definitely required is having the necessary data to hand to make the right decisions and justify them in the eyes of the regulator,” states Wurfbain.
Of the early adopters using Feedstock, Henderson says that some will choose either to increase their fees to investors or enter into an enhanced CSA with their broker – referred to as a Research Payment Account (RPA) methodology, while others will choose to pay for research directly out of their own P&L.
“The RPA model provides all the detail needed for the Article 13 requirements whereas for those paying out of their own P&L, our platform is viewed as a cost-saving device,” says Henderson.
By having more information available, fund managers will be in a stronger position to decide who the best research providers are, for their particular investment needs, and will be able to optimise their research budgets, rotating away from sub-par research providers towards those that are giving their analysts the best bang for buck.
“All of the information we capture allows the buy-side to pay for research correctly and make smarter decisions in the way that they allocate budget,” adds Wurfbain.
He says every fund manager is different and they will each take their own individual approach.
“Whatever methodology is deployed, once you start using Feedstock you become MiFID II-compliant. It is a one-stop-shop for your research commission. It tracks your meetings, what research you are consuming from a qualitative and quantitative perspective etc. We are able to remove a lot of the research headaches that managers will face under this regulation,” says Wurfbain.
Feedstock’s AI system is able to understand what has been researched using behavioural algorithms that analyse the data coming in to the system. It can work out if the research is generic, whether it is from a portal and can then automatically categorise and prioritise the research for each end user.
“In the background, what the AI system does is understand what you value: how much time are you spending on one report relative to another report? Did you make notes on that report? Did you have a meeting with the sell-side research provider? Over time, it builds a comprehensive profile of the data that you like and therefore what you should pay for.
“It learns and grows with you every day that you use it,” explains Henderson.
He says the broader aim is to use the power of AI to better help active fund managers. “If we can create an information channel that automatically understands what is important to the fund manager then it will dramatically increase the efficiency of the active funds industry. We want active managers to use it to complement their day-to-day work and benefit from this technology in much the same way that passive fund managers do.”
One of the inadvertent consequences of MiFID II regulation is that it could potentially allow active managers to harness alpha in terms of the way they use research; not just market alpha, but operational alpha too.
“I think it is a valid point. We provide fund managers with a mechanism by which they can derive more information in a more efficient manner. We’ve spoken to hundreds of buy-side clients and the feedback we’ve had is that they spend, on average, two to three hours, on investment research. On an annualized basis that is about 50 days a year. First and foremost, we recognise that a fund manager’s time is precious. By freeing up their time using Feedstock, it means they have more time to be productive and to focus on running the investment strategy,” says Wurfbain.
In addition, the AI system that Feedstock uses allows managers to compress a large volume of information from which to deduce conclusions and insights without having to rely on manual processes of going through research documents.
“In our view, the productivity benefits are one of the most attractive features of the platform for fund managers and analysts,” concludes Henderson.