March 6, 2015


Assorted targeted spam

You can run, but you can’t hide. Here are a few of the latest things I’ve seen, in no particular order.

  • On a PHPBB-style chat board which I sometimes frequent, there was a thread about do-it-yourself television repair, dormant for over a year. Recently, there was a seemingly robotic post, from a brand new user, that was still on-topic, giving general diagnosis advice and offering to sell parts for TV repair. The spam was actually somewhat germane to the main thread of the discussion. Is it still spam?
  • In my email, I recently got a press release for a local fried chicken franchise celebrating their 40th anniversary. My blogging output generally doesn’t extend to writing restaurant reviews (tempting as that might be), although I do sometimes link to foodie things from Google Reader which will also show up in my public FriendFeed. Spam or not spam?


  1. I’m an admin on a PHPbb-powered forum, and we are constantly fighting spammers there. On the basis of what you’ve described, I’d say it was likely to be spam. We tend to check the IP address of the authors of odd-looking posts against lists of spammy types. Odd post plus dodgy address is almost invariably spam. And I only say “almost” because absolutes are rarely correct.

  2. All depends on how you define spam. If it’s “unsolicited commercial messages”, it’s spam. If it’s “commercial messages I don’t want to see”, then maybe not.

    1) Many people think that blocking some unsolicited commercial messages they would in fact find useful if they saw them is a reasonable price to pay for blocking the ones they wouldn’t want to see. We make those kinds of tradeoffs every day with all kinds of information, or in the physical world by using sharp knives or wearing a seatbelt.

    2) The use of spamming techniques for communication says something (probably something important) about the communicator. So even if a spammed message is facially attractive, the odds that it’s a vehicle for some kind of fraud (commercial, personal, pagerank, whatever) may make it undesirable.

  3. I define spam by “unsolicited”, and not real life relevant. So…
    A friend sending me a lolcat: spam.
    A targeted mail by a online gaming seller that i didn’t agree to give me updates: spam
    My Mom asking me for help with her computer: not spam
    A company that i do not know, sending me an advert for something i really looked for since ages: not spam

    Tho the last one is unlikely for at least another ten years, given how sadly horrible even googles targeted adverts are…

  4. Here,
    Let us distinguish blog spam and email spam.
    1/ email spam is rather simple. legal emails must be double opted-in and provide the ability to unsubscribe. otherwise they are spam
    2/ blog comments: imo, if the comment relates rather closely to the post and comments interestingly (trying), then it is not spam; but a comment with twenty links and no text is spam.
    Pokersavvy Plus Review

  5. The connotation of spam being “commercial” is perhaps too restrictive. Most spammers want to get attention, either on their own behalf or their clients’. That’s certainly true of commercial/fraudulent spam, but the attention getting can go beyond direct profit seeking.

    For example, many office or community environments have “attention economies” of sorts, where the attention one is getting is the currency determining one’s stature and influence in obtaining either financial benefits or getting one’s opinion considered, projects funded, self actualization/validation, asserting one’s significance, etc.

    More often than not, part of the attention game is spamming public communication channels, or more targeted private channels.

  6. Thanks to the fourth commenter (“to spam or not to spam, this is the question”) for providing a perfect demonstration of the problem!

    Much of the spam is related to search engine optimization and marketing. Largely irrelevant spam links, as commenter four provides, do not attract any human clicks, but google and other search engines gladly follow them and raise the rank of the linked sites. There are thousands of hacked wordpress installations out there which post a ton of invisible spam links with every new blog message. You can hire people overseas at very low rates to post comments to any blog or message board containing links to the site you want to raise the rankning of.

    There’s a tragedy of the commons effect where the gaming of the system does not hurt search engines much, but we feel the pain of all of the spam links pretty much everywhere on the internet. Before pagerank, was there link spam?

  7. Andrew S: Good point, I didn’t even realize there was a link at all on first reading.

    But whereas it illustrates the problem in a quite ironic manner, perhaps the very problem statement can be rephrased as “how to (effectively) detect spam that can pass as a genuine contribution on first sight”.

    I’m sure search engines must be trying to address the problem with semantic or stochastic (pseudo semantic) approaches, like correlating linked content to the source subject matter and checking against known spam destinations/topics. But in order for it to work, spam must be reliably identified as such (in the aggregate), by an automated method and without individual human review. Content that gives pause to human readers probably requires semantic reasoning too “deep” for automated classification, when one wants to avoid excluding too many false positives.

    One related problem is that search engine income depends on selling ad placement, which is not that different from spam. The effectiveness of the search results from a search user’s perspective is probably largely of interest to the extent it affects the ad business, i.e. when advertisers complain their ads are not getting the desired response as spammers dilute the search so users are directed away from the “legit” (paid) spam.

    (I hope I’m not stating the painfully obvious.)

  8. Is a legit post with some links at the bottom worse than a post saying “great article” as happens on many blogs?

  9. Spammed message is facially attractive, the odds that it’s a vehicle for some kind of fraud (commercial, personal, pagerank, whatever) may make it undesirable.
    SMO Services

  10. It can be a two-way street in trying to figure out spam. Just about every website is full of links to something, whether useful to one person may not be to another. If it is adding to the conversation, a link included here or there should not be an issue, rather which way the scale is leaning more towards.If you provide helpful information and at the sametime promote your product or service, I wouldn’t consider that as spam because it may be worth it to someoone that helps solve their issue. If we are going to over analyze everything as potential spam, then we shouldn’t even be online and just shut our computer down altogether and read a good book..
    Twitter Buttons