Behind the Scenes in the Deceptive App Wars
All is not well in the app ecosphere. That ecosphere comprises a large number of useful apps that benefit users, and an unknown number of apps that deceive users. The latter are sometimes described potentially unwanted programs, or PUPs.
Both categories need to make money: good apps are upfront with how this is achieved; deceptive apps hide the process.
In recent years there has been an increasing effort to cleanse the ecosphere of deceptive apps. The anti-virus (AV) industry has taken a more aggressive stance in flagging and sometimes removing what it calls PUPs; the Clean Software Alliance (CSA) was founded to help guide app developers away from the dark side; and a new firm, AppEsteem, certifies good apps and calls out bad apps in its 'Deceptor' program.
One name figures throughout: Dennis Batchelder. He is currently president of the AV-dominated Anti-Malware Testing Standards Organization (AMTSO); was a leading light in the formation, and until recently a member of the advisory board, of the CSA; and is the founder of AppEsteem.
But there has been a falling out between the CSA and AppEsteem.
The CSA was officially launched in the Fall of 2015, although it had already been on the drawing board for over a year.
Batchelder was instrumental in getting it started while he was working for Microsoft, where he was director, program management until April 2016.
The CSA was introduced during VB2015 with a joint presentation from Microsoft and Google, demonstrating early support from the industry's big-hitters.
"As a 501(c)(6) nonprofit trade association," writes the CSA on its website, "the CSA works to advance the interests of the software development community through the establishment and enforcement of guidelines, policies and technology tools that balance the software industry's needs while preserving user choice and user control."
In other words, it seeks to develop an app ecosphere where honest developers can be fairly recompensed, via monetization, for their labor. However, it provides very little information on its website. It does not, for example, list the members of the trade association, nor give any indication on how it will enforce its guidelines and policies on recalcitrant apps.
Founded by Batchelder in 2016, AppEsteem is primarily an app certification organization - it certifies clean apps.
However, since a carrot works best when supported by a stick, AppEsteem also calls out those apps it considers to be deceptive and therefore potentially harmful to users.
Batchelder hoped that the CSA and AppEsteem could work together (he was on the advisory board of the former and is president of the latter). The CSA could provide recommendations and industry support on classification criteria, and AppEsteem - at one step removed - could provide the enforcement element apparently missing in the CSA.
AppEsteem maintains what it calls the 'deceptor list'; a list of apps that in its own judgement use deceptive means to increase their monetization potential. At the time of writing, there are more than 300 apps on the deceptor list.
It also actively encourages AV firms to use this list in their own attempts at blocking PUPs.
There is a difficult balance. Deceptive app developers will object to being included on a public shaming list. Apps that get clean need to be removed in a timely fashion.
New methods of deception need to be recognized and included in the bad behavior criteria.
It is, in short, a process wide open for criticism from app developers who are called out.
CSA criticizes AppEsteem
Criticism came last week from an unexpected source - from the CSA. On 10 May 2018, the CSA published a remarkably negative report on AppEsteem's 'deceptor' program titled, CSA Review of AppEsteem Programs. It was, said the CSA, "triggered by a groundswell of complaints and expressions of concern received by the CSA from industry members regarding this program."
The report is largely - although not entirely - negative.
It raises some interesting points. The 'groundswell of complaints' is to be expected; particularly from the apps and the app developers called out for being deceptive.
However, concern over some other elements seem valid. AppEsteem does not seem keen to call out AV products, even when they appear to use 'deceptive' practices (consider, for example, the ease with which the user can download one product and find that McAfee has also been downloaded).
Furthermore, if certification is annual, a certified app could introduce deceptive practices immediately after certification that would go undetected (would effectively be allowed) for 12 months. "There is no more deceptive or risky behavior than that," notes the report.
The CSA report makes four proposals.
AppEsteem should: refocus efforts on certification; work with the CSA to devise consensus-built 'minbar' criteria; balance violator identification and remediation; and embrace oversight and dispute resolution.
'Oversight' implies external management. Refocusing on certification implies abandoning the deceptor app listing. And 'work with the CSA' implies that AppEsteem should take its direction from the CSA.
If not quite a power grab, the report attempts to neutralize the enforcement element of AppEsteem.
AppEsteem's first response was for Batchelder to resign from the CSA advisory board. "I unable to figure out how to remain on the CSA Advisory Board in good conscience," he wrote to the CSA. "Which sucks, as I've pushed for CSA to get operational and remain relevant, sent potential members its way, and worked hard to help it succeed. But being an advisor of an incorporator-status organization who is conducting a 'confidential' investigation into AppEsteem's certification program without involving AppEsteem makes no sense at all."
AppEsteem's second response was to establish CleanApps.org; which is effectively an alternative to the CSA. "AppEsteem needs CSA," comments one source who asked to be anonymous, "or at least some organization that can provide guidelines and some kind of oversight of what AppEsteem is doing... It seems that this new player is in fact a company created by Dennis trying to get rid of CSA."
That partly makes sense.
If AppEsteem cannot work with the CSA, it must find a similar organization it can work with. "After I disengaged from CSA, Batchelder told SecurityWeek, "we realized that AppEsteem had to find a way to get the vendor voice and to reassure them that we're doing things fairly (the stuff we had hoped CSA would do). So, I incorporated CleanApps.org and recruited its first board from some of our customers (I know, it's like a soap opera), and then resigned/handed it over once the board launched. Our goal is that once CleanApps.org launches, we'll give them insight into our operations."
To the CSA, he wrote in February, "I wanted to let you know that we have determined that it's in best interests of both ourselves, our customers, and the vendor community if we had oversight and a 'voice' specifically representing the vendor community...
We won't become a member or hold any position in CleanApps.org; they will self-govern." (He has since made it clear that he does not mean 'oversight' in any controlling manner.)
AppEsteem's position seems to be that the app ecosphere requires three organizations: AppEsteem to enforce good behavior among the app developers; the CSA to represent the market in which apps operate; and CleanApps to represent the apps and app developers.
But it is clearly concerned over the current relevance of the CSA. "I think the biggest hole with CSA," Batchelder told SecurityWeek, "is that they never finished forming: it's still just... as the only member, and what we felt was that when [that member] had an issue with us, CSA went negative... it's problematic to us that they're not formed after four years."
If AppEsteem needs something like the CSA to be effective, the CSA needs something like AppEsteem to be relevant.
AppEsteem's third response is a short blog posted on the same day as CSA published its report - Thursday, 10 May 2018. There is no indication of any rapprochement with the CSA. "But we also want to be clear," writes the author: "if you think it's fine to treat consumers as exploitable targets for deceptive and aggressive software, we totally understand your desire for us to leave you alone. We strongly suggest you either get on board or find something else to do with your time, as we're going to continue to tune our Deceptor program to find even more effective ways to disrupt your ability to hurt consumers."
The way forward
It is hard to see how any outright deceptive app produced by developers simply out to get as much money as possible will ever be persuaded by force of argument alone to abandon deceptive practices.
This seems to be the approach of the CSA; and it appears - on the evidence of its website - to have achieved little in its three to four years of existence.
Indeed, the one and only report the CSA has published is the report criticizing AppEsteem. Before that, the previous publication seems to be 'update #7', probably written around March 2016.
If the CSA has achieved anything, it is not saying so. At the very least, it could be urged to be more transparent in its operations and achievements - even a list of members would be useful.
Meantime, if the new CleanApps.org gathers pace and support, the CSA itself will become increasingly irrelevant in the battle against deceptive apps; that is, potentially unwanted programs.