For as long as people have been doing credit analysis, analysts have been thinking about how to avoid spending hours manually entering numbers into spreadsheets to build financial models. The offshore analyst model, part of the wider Knowledge Process Outsourcing industry, was meant to be the solution, but for the majority of people it hasn’t delivered the benefits promised. Technology solutions are now stepping into the mix to address this.
To form your own opinion on a credit, you have to analyze a company’s fundamentals. You don’t just look at the most recent set of numbers, but the results over time, the notes to the accounts, the annual presentations, etc. After building a financial model you then need to do your own analysis to draw out insights and form your conclusions. End-to-end, it is a very labor-intensive and time-consuming process.
Most businesses would prefer to avoid paying expensive employees to do low value work, and the credit investment industry is no different. Credit investment firms don’t want to pay highly skilled, expensive analysts to perform data entry tasks. Similarly, most credit analysts don’t want to spend their day punching numbers into spreadsheets, either. This isn’t a new problem, and the prevailing solutions have been to hire more junior analysts to crunch the numbers or to outsource model building to a lower cost provider.
Knowledge Process Outsourcing (KPO) has been used by credit investors for decades in an attempt to solve this problem and amplify their analysts’ work. Armies of highly skilled modelers in lower-cost countries have been employed to extract information from financial reports and transpose the information to spreadsheets to form the basis for financial models and analysis. For those in Western Europe and the US, the outsourcing model has had the further advantage of being based in a different time zone, enabling overnight turnaround of information. It is an appealing proposition, in theory.
In reality, few actually achieve what they are intending to with KPO. Why? Because KPO solves one of the problems of credit analysis - swapping cheaper labor into the process - but it introduces new problems and doesn’t address others. We’ve previously worked with KPO firms and have spoken to a number of buy-side and sell-side analysts who have all worked extensively with KPOs, too.
“We tried several different arrangements, but the result was always the same – lots of frustration on both sides because the outsourced teams were never really in sync with the needs of our business. Their output often needed to be re-worked, causing us to question what we were actually paying for.”
One of these analysts made the following comment, reflecting on 5+ years working with two different offshoring suppliers: “We tried several different arrangements, but the result was always the same – lots of frustration on both sides because the outsourced teams were never really in sync with the needs of our business. Their output often needed to be re-worked, causing us to question what we were actually paying for.”
Analyst criticism of outsourced model building tends to fall into 5 categories that anyone with experience of using KPOs can relate to:
Humans make mistakes when manually entering numbers into spreadsheets. That’s an unavoidable fact of data entry and one that people accept. But when you outsource model building, how do you verify the numbers in order to be 100% certain you can make an investment decision from them? A typical process to address this challenge entails checking the first few models with a fine-toothed comb, spotting any errors and subsequently providing iterative feedback to the KPO firm. The assumption is that if you set the bar high, no more errors will come through. Unfortunately, life is not so simple, and the analysts we spoke to mentioned a constant nagging fear that they were going to get called out by their boss on a bad trade that was based on an error in a fundamental model that they hadn’t built themselves.
It’s not just inaccurate models that you are faced with from your KPO partner. An equally challenging issue is the variability of models you receive depending on the exact KPO employee doing the work. No matter what incentives are used to try to align goals, nor how templated you try to make the process, everyone has their own way of building models, interpreting information and formatting spreadsheets. The result is models that come back to you just different enough to require spending time re-engineering in order to get them exactly right. A few hours re-engineering every credit soon adds up to a lot of days and weeks.
Most off-shoring services are based at least 5-6 hours away from the location of the client. For more time sensitive matters this setup creates a host of its own problems, especially around both model updates and creation. Taking each in turn, even when office hours overlap, ensuring seamless communication with respect to the update of new numbers is challenging at best: has the analyst seen the results, are they at their desk, have the models been updated (correctly)? This back-and-forth is difficult enough for a few new reporters, let alone a busy pre-market open. For companies that report after the market close, depending on the set-up of the offshore service, clients run the risk that models are not updated until the following morning, defeating much of the purpose of the offering.
In order to achieve the kind of market coverage that firms are looking for, the minimum team size of the people I spoke with was 3, the maximum over 10. Even with a KPO manager as part of the package, you’ve added a bunch of remote, 3rd party “employees” to your reporting line. Managing a team is hard. Managing a big team is harder. Managing a big team in a different country and different time zone and with a different first language… a tough job just got way tougher.
Knowledge Process Outsourcing was once attractive because of the labor price arbitrage. Skilled labor in lower cost countries was, seemingly, plentiful and the difference between lower cost and higher cost regions was substantial. Fast forward 15 years and 15 - 25% compounded annual salary inflation for the most skilled workers makes that arbitrage much less attractive. You can spend hundreds of thousands of dollars outsourcing model building, only for the individual analysts in your credit team to ignore the model provided and start completely from scratch, defeating the whole point. It’s never as cheap as you hope. In fact, it’s not cheap at all these days.
You can spend hundreds of thousands of dollars outsourcing model building, only for the individual analysts in your credit team to ignore the model provided and start completely from scratch, defeating the whole point. It’s never as cheap as you hope. In fact, it’s not cheap at all these days.
Where once KPO was the solution, more and more credit firms today are looking to technology to deliver efficient, accurate financial modelling services. After all, replacing repetitive and time consuming tasks is what machines were built to do. Technology applied to credit markets allows for a trust in the data, speed of results and transparency through to source documentation that KPO solutions could never provide.
Machines aren’t humans though, so when thinking about whether a technology partner can deliver better results than the KPO model, the desired outcome needs to be clear.
Credit Analysts want access to fundamental company level information on which they can form their own opinion and ultimately base investment decisions. They want the information in a consistent format, in a timely fashion, and for it to be easy to reconcile.
In the credit analysis process, what you gain in accuracy, automation and lower management costs, you worry that you’ll lose in customizations, flexibility, and customer service. These are natural concerns, after all, you have a computer doing the task for you rather than a human. That’s why choosing the right technology partner is so important. You should be looking for a technology partner that delivers your desired outcome, with minimal costs.
How do you minimize the costs and maximize the benefits?
When analyzing a credit, you like things done and presented in a certain way. Some technology providers (and the product teams in those technology firms) have a preference for offering the Henry Ford level of customizations: “you can have any color you want, provided it’s black”. But just as cars today aren’t all one color, so too do technology firms offer more customizations and personalizations once they have moved past the Minimum Viable Product stage.
How does your technology solution present company reported financials - do they provide a clean read from the original statements or is it a sea of subjective adjustments? Does your technology solution allow you to edit and save a model with your own view of FCF, seamlessly updated for new reports - or are you stuck with the original model? Does your technology solution utilize all the data points in its system to provide a top-down Comparables view of the market based on fundamental credit data?
These are just a few points to consider when picking the right technology partner.
You’re not going to find a technology partner whose coverage universe perfectly matches your investable universe. The off-the-run opportunities that no one covers will always exist, but those opportunities are not the problem you are solving for. You need to choose a partner that covers a significant portion of your available, investable market. Market coverage counts. Your provider also needs to be reactive to requests for new coverage. After all, new names come to market all the time and you need to be confident that those new opportunities will be added in an expedited time frame.
You need to choose a partner that covers a significant portion of your available, investable market. Market coverage counts. Your provider also needs to be reactive to requests for new coverage.
How many times have you been sold the software dream, signed up to a new provider, and then been left high and dry when you come to use it. “Your call is important to us, but we are experiencing unusually high call volumes”? It’s a big concern for buyers of any Saas solution: will you actually receive expedited support when required. The solution is obvious: make sure you choose a technology partner committed to service and support, one with a proven track record of providing it.
Over the past 20 years, the offshore analyst model has become a common approach for firms looking for cost-effective data entry. But while the benefits sound amazing in theory, in practice the results do not measure up. Automated data extraction is increasingly being utilized in many parts of financial markets, and many credit investors now view this technology as a more powerful, flexible solution to supporting their model building requirements.
At Cognitive Credit, we’ve developed technology that specifically extracts earnings data and delivers ready-to-use credit models - for over 1,500 issuers - in minutes.
To find out how we can help transform your credit investment business, get in touch today.