In an effort to provide this useful tool to a larger community, we recently exposed our search engine outside the password-protected firewall. This can be accessed at http://search.gridstoneresearch.com.

While this is in beta, we have been introducing new and interesting features to this already unique search engine. For example, today we’ve introduced a feature that shows a tag cloud of the tickers that matched a user’s search term. This visual presentation provides a very intuitive view into the distribution of commentary between various companies.

For example, to see who was really impacted by the war in Iraq, I ran a search for “Iraq”, filtered my results to exclude 10-Ks/Qs (a lot of companies list Iraq in their risk factors) and got this tag cloud as a result. The Government services companies (KBR, MANT, HRS, LLL, LMT, etc.) were the obvious candidates for such a search, but I was surprised to see companies like OSK and RDY in the mix. Another useful way in which one can see what’s impacting various companies.

I co-founded Gridstone 3 ½ years ago, and have spent most of that time building out our product platform. To spearhead our market development efforts, I recently relocated from the San Francisco Bay Area to New York, where a majority of our customers are.

With the support of a few key strategic customers, we have spent the last 3 ½ years building a unique and valuable set of capabilities along many different dimensions. While that process is never complete, we have reached a point where the time has come to aggressively take these to market – by ourselves and more importantly, with partners. That is going to be our focus going forward – to expand our relationships with our current customers by anticipating and serving their research and data needs better, working with partners to penetrate the market, and leverage our capabilities in new and exciting areas.

I thank you for your support thus far, and look forward to your continued inputs and suggestions as we enter an exciting new phase of growth.

PR Ganapathy

There’s a Vanguard ad running in the print media that I quite like. It visually makes the point very eloquently that management fees matter. With the market performance of last year, Vanguard probably thinks that the iron is hot for passively managed funds to strike and gain dominance. I wrote about this shift in my last post.

Investment costs are becoming important in the hedge fund industry too. With uncertain markets and lower leverage to goose the returns, LPs are asking tough questions about the 2/20 hedge fund economics. Mutual funds may not have to change their economics but few have been spared the twin blows of declining prices and redemptions.

Our message to the asset management industry is not very far from what’s depicted in the picture above. “Investment Research Costs Count.” is what we would have said. But that’s not all. Relying on a technology enabled service for Research Data Analytics relieves your analysts from the tedium of pulling data from filings to assemble a spreadsheet. And the final outcome is higher quality, auditable and with a faster turnaround time.

Better, Cheaper, Faster. That’s us. Talk to us.

There is no question in my mind that the share of equity assets will shift towards passively managed funds. But I can’t agree that the shift will be as dramatic as that. Passive and actively managed strategies don’t just compete with each other for market share. They also affect each other’s success. Here’s how.

A market that is dominated by active strategies, makes it very difficult for these competing active strategies to make money, especially compared to the broad market. In such a market, the lower costs of passively managed funds can make them attractive. That’s the kind of market we have right now.

On the other hand if we have a market where passive investing dominates completely, it should be easier for a few active strategies to make money because there are very few people actively interpreting new information to assess its impact on companies.

In effect, there should be a stable state in the market where the share of active funds is not so large that they all compete away their informational advantage over passive strategies to an extent that they can’t justify their fees. This thesis borrows loosely from the work of biologist Maynard Smith who applied this to evolutionary theory. Of course, this won’t help us predict what the share of each strategy is in the theoretical stable state, but my hunch is that it is not close to either 90/10 or 10/90.

The real value of a researcher’s work is in bringing his or her insight about the company to bear upon its analysis. Yet they spend an inordinate amount of time on managing historical data and quarterly updates in spreadsheets even though they know that this is not the best use of their time. What if this was not so? What if everything related to Research Data Analytics were no longer the researcher’s headache? They would still control the design of the analytics, but wouldn’t have to labor over its construction. What would it take to make this so?

The Story of an Earnings Model

Let’s take the example of an earnings model – the bread and butter of research in a typical equities mutual fund. It contains historical time-series data on the company’s financials. Based upon certain proprietary assumptions, the analyst projects financials for the company into the future in order to ascertain the intrinsic value of the company.

Building a ‘bottom-up’ earnings model is an arduous task that can take a week or two of an analyst’s time to complete. And then you have to keep it updated! What if you were an analyst and wanted to be able to get an earnings model built to your specifications with just a couple of hours’ work from your end? Where are the hurdles to making this happen?

The first hurdle is the data. Any bottom-up earnings model will use operational data, breakouts, one-time items etc. in addition to data from the three financial statements. You will have to get that data yourself from the filings (or your Associate will, if you are lucky enough to have one). Depending upon how much history you want, it could take days just to assemble the data.

Even if you somehow assemble the data, you still have to update the model manually. There’s no getting around that. Every quarter, like clockwork the company reports new data. And sometimes restatements, which just means more data entry.

The third hurdle is customization. You will want the model built just so. “The quarters should go from right to left. The balance sheet and cash flow statement should be on separate tabs. Revenue must be driven by units shipped times ASP.” That’s the way you want it. Nothing else will work. This makes a model a very personal thing and something that is impossible to get right by borrowing or buying models off-the-shelf.

The fourth hurdle is quality. When there is so much manual work involved, there are bound to be errors. Errors can be very costly in this business. They could be in the data or in the spreadsheet formulas. That’s why you have to check and recheck the model several times when you build them yourselves. When someone else builds it for you, it gets tricky - do you trust the data or go check every data point? Which defeats the whole purpose, doesn’t it?

A new way of thinking about Earnings Models

Are these hurdles insurmountable, then? Can Earnings Models never be provided as a service to an analyst? Is he forever condemned to building them himself?

Not necessarily. A clever mix of services supported by the right technology can be just the ticket.

Gridstone Research has developed technology that greatly automates the extraction of data from filings and then allows a user to pull it directly into Excel. Gridstone Associates can use this research platform to build models very easily and quickly to a client’s specifications. Clients can either send us a model for a similar company that we can use as a template. Or they can answer a brief questionnaire about how the model should be laid out and how projections should be made. Total time spent by client should be less than an hour if the specs given are reasonably complete. Another iteration with the finished model to dot the i’s and you have an earnings model built exactly to your spec.

The earnings models can be automatically updated if the client uses our Excel Add-in. When the next quarter rolls around, the new data automatically pops into the spreadsheet. All data is auditable – two clicks will show any data point highlighted in the filing. The assumptions for the future projections are kept on a separate tab so that you can change them without having to change formulas.

Pricing is by the model, fixed and completely transparent. Based upon a couple of variables you should be able to compute the cost of the model yourself, well before you spend any time explaining how you would like the model built.

On Demand Earnings Models

Should model-building done the Gridstone way be called a Utility? Not really. Unlike a utility, it is highly customized to exactly meet client requirements. But in some other important ways, it is close enough that we call it On Demand.

• The quality of the deliverable is uniformly high.
• The demands on your time are minimized.
• The price is transparent and parameterized.

On Demand Earnings Models. This unique service is available from Gridstone right now.

Photo by Joe Shlabotnik

Analysts of all hues build Excel spreadsheets to analyze companies. Some of these spreadsheets can be very complex. Most of them will require manual data entry which can be a mind-numbing exercise of downloading filings and copy-pasting data from the filings into the spreadsheet.

Gridstone Associates use our proprietary research platform in providing services to its clients which makes the job of building such spreadsheets all brain and no brawn.

• Deep data on the company is pre-extracted from the filings using our data extraction technology. All time-series data – financial, breakout, operational and footnote data – is available to our Associate through our research platform that includes an Excel Add-in.
• The easy-to-use Excel Add-in allows a Gridstone Associate to download fundamental data into Excel directly, in any layout desired.
• The Excel spreadsheets constructed can be kept linked so that future updates are automatic.
• Every data point is source-linked back to the filing that it came from. In two clicks you can see the data point highlighted in its filing – all without leaving Excel.

Depending upon the service contracted for, the Gridstone platform can also be used by clients. Here is a short video that demonstrates the Excel Add-in.

http://www.gridstoneresearch.com/excel-add-in-video/

Stratify’s vision of “making Unstructured Data actionable” could very well have been written for us. Stratify uses its text analytics technology to provide industry leading EDD services to the legal profession. Gridstone’s goal is to make company research powerful, easy and affordable by harnessing similar technologies. Ramana understands the disruptive power of such technologies at a level that very few do.

As an entrepreneur, Ramana built a scalable, successful business after safely navigating the dotcom bust. His advice and counsel are going to greatly benefit Gridstone.

It’s great to have Ramana in our corner.

Tell us about your firm and the role you are in

I work as an energy analyst for The Mitchell Group, a long-only energy focused investment group with roughly $750mm AUM. Our investment style can be best described as Value, supported by fundamental, bottom-up analysis. Our investment time horizons are in the 3 to 5 year range and so we have a very low turnover. We tend to invest in theses supported by valuations and fundamentals.

Describe your own research process

My coverage universe includes various sub-segments within the energy sector: alternatives, coal, refineries, oil and gas, equipment providers, service providers, offshore drillers, land drillers, and exploration and production. Just within the oil and gas exploration and production sub-sector, I follow about 50 companies and maintain models on about 25. I will use models to screen for value and the team will discuss those ideas and conduct meetings with company management as part of our due diligence.

How does Gridstone help you in your research?

One of the biggest time sinks in the modeling process is the manual data entry and verification. For sector focused, small investment groups like TMG, analyst time is a precious commodity and as such, tools like Gridstone become invaluable. With Gridstone, data from SEC filings and transcripts can be mapped and linked to my models. The data entry process becomes automated, thus freeing up time for analysis.