The Problem With Robot-Written Earnings Report Articles

Includes: KTOS
by: Patrick Connolly


Associated Press's Automated Insights service automatically generates around 3000 articles per quarter giving a summary of company earnings reports.

The case of Kratos's (KTOS) recent earnings reports demonstrates a big difference in sentiment between the information in the press releases and the picture painted by an automatically generated article.

Readers are suggested to be aware of when a report is automatically generated in order to avoid developing a distorted 'first impression' of a stock.

I am long a tiny amount of Kratos (NASDAQ:KTOS) stock. I also recently wrote an article here on Seeking Alpha about the benefits this company might have from the imminent growth of the global military laser market. I published it a few days before the most recent earnings report which took place on the 12th of March (the transcript of which is available here), so the fundamentals there were based on the 3rd quarter from last year. For these reasons, I was somewhat attentive to the 4th quarter earnings report. The report was mostly good news, enough to take the stock up well over 8% the following day (though the gain was largely reversed the next day in yet another broad market sell-off).

Some background on Kratos, if you don't know about it: this workforce of 3600 mostly technical staff are largely located within US Military bases. The company receives constant orders for defense technology (about 70-80% of revenue), as well as orders for public service projects. The company's focus is on developing a diverse array of defense technologies including command, control, communications, computing, combat systems, intelligence, satellite systems, unmanned systems, electronic warfare and much more. The company most typically takes smaller contracts to develop key technologies, but also collaborates with larger contractors on key projects as well. The company has been acquiring other companies and paying debt from these purchases is one reason earnings per share have remained in negative territory. However, those long Kratos have been waiting for the company's tech lead to result in a growing position in the defense market, as well as growth from acquisitions beginning to show. The company appears to have been reporting more and more contracts of late, and shareholders had hoped for a sign of this positive move in the latest earnings report.

For those who don't know the results, Kratos appeared to show this sign, beat earnings estimates, with their $0.11 adjusted EPS beating estimates by $0.04 (including large interest payments on debts related to acquisitions). Generated operating cash flow was up at $25M. The company recorded an impressive total of $320M for new awarded contracts and bookings during the quarter, saw an increase of funded backlog of $129M (now at over $660M), with total backlog at $1.1B as of the end of the quarter. The qualified bid pipeline was reported at $7.4B.

The company has usually reported according to two segments: KGS (defense) and PSS (public service). However, in the present earnings report, they indicated they would separate a US (unmanned system) segment out from the KGS segment in future due to the significant expected growth in the US segment over the next two years. In this regard, though the US segment is viewed as a key growth segment for the company, it saw a reduction in revenue from the previous quarter of $1.8M from $22.9M to $21.1M, reportedly due to completion of a project, the next stage of which begins in 2015 (many new orders are reported for this segment going into 2015). On the other hand, the KGS segment showed sequential growth in every quarter of 2014, the final quarter being no exception, going up to $159.3M from $151.4M the previous quarter. The PSS segment also shrunk $1.7M from $42.8M in the third quarter to $41.1M in this one. The company attributes this partly to delays in contracts but also partly because the company is intentionally moving away from less profitable contracts in this segment towards more profitable contracts.

The market jumped on the news, and champagne corks were already popped, when an article appeared online having much the same effect on my party mood that a blunt instrument would have (available on CNBC here). The information reported by this short article, in total, is that Kratos reported a quarterly loss of $2.2 million (a loss of 4 cents per share), and that the revenue of $221.5M missed forecast by about $10.2M, and that the 2014 year revenue was $868M. The yearly loss was reported as up from the previous year, sitting at $78M or $1.35 per share. Finally, though it was indicated that the share price high of $5.93 for the day was up 18% since the beginning of the year, it was still 22% down over the last 12 months. That's it.

The article painted such a bleak picture of the stock, and left out so much of the information that would show how the company is getting where it is aiming to get (in the face of ongoing contractions of defense spending), that at first I thought it appeared a shameless attempt to spark selling by those shorting it (and no doubt appalled by the post-earnings rise). It was only then that I check the authorship: 'Associated Press, Automated Insights.'

You may want to check out this article from the Verge (here). Automated Insights saves manpower by generating automated articles summarizing companies' earnings reports, about 3000 per quarter. 'Robot-generated' articles are written in a standard AP style that appears at first glance to have a human writer, though not a particularly imaginative or verbose one. Of these, about 120 receive an added review by a human, and a few very select companies receive an entirely human review. It is suggested that this does not reduce AP's spending on human writers, but rather allows for a greater number of reports to be published, so good for AP.

The point I'm trying to make here is not whether you should be confident or not about the company's success. Concern has been expressed about declining defense spending which has been a reality for contractors in the last few years (though Kratos product diversity has provided some resilience in the face of this). Concern has also been expressed about the fact that Kratos has a huge amount of debt that keeps turning profits into losses, though the point can be made that the debt incurred (through acquisition of other companies amongst others) represents growth spending that will pay off later.

Rather, the point to made here is that the metrics selected for reporting by the automated process are not necessarily the most important for showing where the company is or is not, and may present a very distorted image to an unwary reader.

There are no doubt those who would suggest that the metrics reported by the 'robot' article may be useful as an objective (and sobering) reminder of the bigger picture, a surprising antidote to the spinning present in 'human-written' articles. Certainly the information in it should anyway be considered by anyone investigating the company for themselves. However, while I have selected the case of a 'robot-written' article that may darken a potentially good play, the reverse could hold just as well. An automated article may well present information that looks very good, when in fact there may be serious concerns regarding a company's position and outlook.

Of course, anyone serious about investing their money will go on to do a lot more reading than one article (though psychologists tell us that first impressions do last), so the matter may not be anything more than nuisance-level. However, I know that in future, I will be scanning the by-line of earnings reports summary articles more promptly to protect me from any unwarranted shocks (positive or negative).

Of course, the robot may indeed have been shorting the stock, they are getting so very clever these days.

Disclosure: The author is long KTOS.

The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: This article was written by a human. Where not referenced, the information contained in this article was sourced from both Google and Yahoo Finance, as well as Kratos's own site.