Digital Financial Reporting

Maybe this won't put the Fortune 500 list out of business any time soon (or heck, maybe it will) but here is a prototype list of the top 1000 SEC filers:

• By Assets
• By Revenues run rate
• By Net Income run rate
• By Net Cash Flows run rate
• For good measure, here is the current Fortune 500

Let me point out a few things.  First, keep the word PROTOTYPE in your mind. Not saying these are precisely right just yet.  They seem to be, but I have to debug my algorithms.  This prototype is intended to help people see the power of XBRL, not be the definitive source of the top 1000 companies.  One day maybe.

Second, my prototype uses only the 10-Q and 10-K filings between 2011-04-01 and 2011-08-19.

Third, note that I am using the run rates for the information which relates to a period of time.  This is because the 10-K information is for an entire fiscal year, whereas the 10-Q information is for one or more quarters.  To make this comparable, I am dividing the number by the days to get the amount per day, which is called a run rate.  Note that the Fortune 500 is updated yearly.  I can update this every time a new filing is received by the SEC theoretically.

Fourth, I could add other categories: total R&D, net operating cash flows from continuing operations, yada, yada, yada.  Why?  Space on the internet is free and because once the query is written, running the query is free!

Fifth, I could provide the top 5, top 10, top 25, top 100, top 500, top 1000, top 2500, top 5000, or heck every filer!  I could do it by SIC code, by home state of the corporation; as all that metadata is out there to use.  Want a different sort?  Fine, add more metadata.

Finally, this type of functionality is not limited to the SEC XBRL financial filings.  Any data set expressed in XBRL can be accessed just as easily.  I am using Microsoft Access, no XBRL processor, and I am not even a very good programmer!  Granted, the queries are not that challenging and my queries are limited.  Imagine what you can do with an XBRL processor, with a more powerful database, and using more sophisticated queries.

P....R.....O.....T....O.....T.....Y....P....E!!!  Don't beat me up if I have errors just yet. I am working on tuning this. The point is to show the possibilities, not analyze companies.

This is one of the more interesting analysis reports that I have put together, a summary of the concept used to report revenues by a filer, sorting the filers by SIC.

So here is the synopsis.  Of the 5525 SEC XBRL filings I am looking at, 4538 (82%) used one of these 10 concepts to report revenue:

• us-gaap:Revenues (1754 which is 32% of all filers)
• us-gaap:SalesRevenueNet (1446 which is 26%)
• us-gaap:InterestAndDividendIncomeOperating (554 which is 10%)
• us-gaap:SalesRevenueServicesNet (336 which is 6%)
• us-gaap:SalesRevenueGoodsNet (324 which is 6%)
• us-gaap:RealEstateRevenueNet (66 which is 1%)
• us-gaap:OilAndGasRevenue (62 which is 1%)
• us-gaap:RegulatedAndUnregulatedOperatingRevenue (18 which is <1%)
• us-gaap:HealthCareOrganizationRevenue (14 which is <1%)
• us-gaap:RevenueMineralSales (9 which is <1%)

That leaves 942 filers which either used some other concept, created an extension concept, or did not have any revenues.

So, the fact that one third of all filers used "us-gaap:Revenues" makes an important point.  Are all those more detailed breakdowns of revenue more helpful or does it cause problems suchs as having to figure out WHICH concept was used to report revenue.  By contrast, imagine if every industry had their own concept for assets or broke down assets by industry; us-gaap:RealEstateAssets, us-gaap:OilAndGasAssets, us-gaap:RegulatedAndUnregulatedOperationsAssets, us-gaap:HealthCareOrganizationAssets.  Is that helpful or unhelpful?

If you go through the breakdown of which concept was used by industry (the SIC), one would expect to see some patterns of usage of specific concepts by industry.  But I don't really see any real patterns that stand out.

For example, go look at SIC 1311 CRUDE PETROLEUM & NATURAL GAS.

• us-gaap:Revenues (106, should the more specific concept us-gaap:OilAndGasRevenue have been used?)
• us-gaap:OilAndGasRevenue (33 should the more general concept us-gaap:Revenues have been used, clearly analysts understand what industry the company is in)
• us-gaap:SalesRevenueNet (18)
• us-gaap:SalesRevenueServicesNet (4)
• Other concept used or had no revenue  (34)

I predict that filers will look at the concepts being used by other filers and they will likely make adjustments to which concept they are reporting.  Perhaps the SEC may provide additional guidance to create more consistency, perhaps not. Software algorithms can certainly figure this stuff out and give those analyzing the data what they need, but writing those algorithms will be costly and therefore not everyone will benefit from those adjusted sets of information.

From what I have seen, the reporting of revenues has more variability than any other area of an SEC XBRL financial filing that I have looked at.  I would suspect cost of revenues to be in the same boat.

Someone made me aware of the WICI EBR/Gartner KPI Initiative.  Their web site provides the following description of this project:

The EBRC/Gartner KPI Initiative is a market collaboration to develop standardized key performance indicators initially for the High Technology, Insurance and Retail and Consumer Goods sectors as extensions to the WICI framework. Gartner will facilitate the development of the industry sector standardized KPI concepts via private client groups to include leading company and investor/analyst participants. It is expected that the collaborative KPI effort will take approximately six months from inception.

They appear to have an XBRL taxonomy for the standardized set of KPI concepts.

Someone asked my if I could provide this information for two specific SIC codes.  So what I did was provide the information by SIC.  Here is another version of the research and development information. In this version you get a list by SIC of research and development expenses, total assets, and the percentage of the two; plus you can drill down into the SIC code to see a list of the SEC filings which make up the list.

Also, I aggregated these demos here. Keep watching, I may add some more.

This is starting to get fun! I did another experiment, grabbing the concept us-gaap:ResearchAndDevelopmentExpense from my set of test filings.  Of the 5525 filings, I got 1236 hits (22%). So for example, Microsoft is #1 in R&D Expense spending about $9 Billion between July 1, 2010 and June 30, 2011 (which is about 8% of total assets) per their 10-K filed July 28, 2011. Second is Oracle which spends about half of what Microsoft did for a similar period of time.

If anyone has any query ideas, send me an email and I will see if I can grab the information.