Poll: SPAM - do Plusnet owe users an apology?

[#1]

As most people are aware Plusnet are in the process of implementing a new anti-spam system using Postini. However on 11th December another anti-spam system which was on trial (Criticalpath) was removed from service (Service Status announcement). Ever since then many users have seen significantly higher levels of spam getting through.
There has also been an issue with the identification of emails as [-SPAM-]. In spite of warnings from users that DSPAM was picking up the headers added by Postini for those on the trial - training continued for several weeks until something was finally done about this earlier this week. This has led to a very significant increase in the number of emails falsely being identified as spam.
In addition, in spite of users faithfully reporting the missed spam and wrong identified ham, for a week over the Christmas holidays, no training was done of the spam filters (it is supposed to be a daily task).
I think Plusnet has badly failed it's users over the last month. There seems to have been a degree of complacency that Postini is the solution, and it doesn't matter what pain the users have to suffer in the meantime. The results Postini trialists are reporting are very encouraging - but that is little comfort to users not on the trial who have suffered overflowing mailboxes.
Do you think an apology is called for?

[#2]

I have voted yes for PlusNet to apologise for the way they have dealt with the matters surrounding the Postini implementation.
Of course an apology inherently carries with it, the promise that special efforts will be made to ensure that a similar debacle does not occur.
This is of course down to senior management. (Where are you Neil Armstrong?)

[#3]

Seeing as we are appearing to speak our minds here.......
Although I agree with much of what you say John, having a non Postini business account with all the resulting daily clean-up work, I don't really see that a 'public' apology would achieve anything apart from appeasing a few.
Considering the vast majority of customers don't appear to know that the community site exists never mind read the forum, where would you suggest the apology is published.
Hindsight is a wonderful thing but I'm not totally singing the praises of PN on the whole Spam issue as I do believe it could have been handled better.
The removal of Criticalpath has caused difficulties for many customers and it's a pity that couldn't have been timed better.
I respectfully abstain from voting as I feel an absolute vote either way would need quantifying.

[#4]

Quote from: Strat

Considering the vast majority of customers don't appear to know that the community site exists never mind read the forum, where would you suggest the apology is published.

Firstly I'd like Plusnet to admit in here that they have messed up with this.
Secondly I'd hope to see something in the next newsletter sent to all users (and not just an article trumpeting how good Postini is which is what I'm expecting will appear!).

[#5]

Possibly if a collation of all 'admissions of omission' by various PN staff across the Forum relating to Spam was performed, an apology would emerge (I'm not inclined to do it).
However John, taking your reply into consideration and in support of those less informed I mentioned earlier, I have changed my vote to yes.

[#6]

I've voted 'No' because I think that, on the whole, Plusnet have tried their hardest to cope with the spam situation.
Regarding the shortfalls with DSPAM, they were/are never going to get the 'long word' spam detected using DSPAM because DSPAM only looks at whole words - not parts of words. All of the long words used are different.
On the Postini trial front, it was always going to have teething troubles. No system is perfect and the majority of problems for Postini trialists have actually been due to DSPAM which, with hindsight, they might have picked up on a little sooner.
The Critical Path trial had to be ended because the allowed trial period had ended and, unless it was to be purchased, they had to switch it off. If they had purchased it Jelv, you would not be looking forward (with such relish) to getting Postini implemented.
So, I think that their hands have been tied by circumstance and smart spammers and therefore it is not Plusnet's fault.
Tony

[#7]

Just a thought.
People are talking a bit about 'hindsight'.
What of course must be remembered that most of the set-up and testing procedures used by PlusNet should not be new.
It is not new at all to make sure that a testing method should adequately follow the actual working conditions. The 'inhouse' testing did not follow the actual conditions encountered by the customers.
There will always be problems if a testing procedure is flawed.
There is an unfortunate habit in the 'computer' world to get the customers to test something. A lot of the customers seem to feel proud and important in being involved.
This is strange to me because I beleive that the 'seller' has the responsibility to make sure that their product is fit to use.
Is there not a law which says something like that.
To summarise, what is being called 'hindsight' are well established principles and have been established for a very long time.
I'm all for making sure that if I test something I test it properly.
When I agreed to take part in the Postini 'trial', I never dreamed that the testing had not been devised adequately.
But is it true that the folk who suffered most are the people who did not take part in the trial, because the methods used were faulty.

[#8]

Quote from: jelv

In addition, in spite of users faithfully reporting the missed spam and wrong identified ham, for a week over the Christmas holidays, no training was done of the spam filters (it is supposed to be a daily task). I think Plusnet has badly failed it's users over the last month. There seems to have been a degree of complacency that Postini is the solution, and it doesn't matter what pain the users have to suffer in the meantime. The results Postini trialists are reporting are very encouraging - but that is little comfort to users not on the trial who have suffered overflowing mailboxes.

Agree 110%. Apologise and do whatever you have to do internally to improve the way you manage these situations.

[#9]

I don't want an apology - I just want them to fix the problem, or tell us how to fix it ourselves.
I never had a real spam problem with plusnet before the infamous data breach last spring:
http://www.theregister.co.uk/2007/05/15/plusnet_spam_attack_may/
I only get spam to email boxes that were created *before* that date, so I assume it mostly stems from that period.
I've been with plusnet a long time now, and though I'm generally happy with the service, they have a very sorry history when it comes to managing email:
http://search.theregister.co.uk/?q=plusnet+email
Apologies if I'm going over old ground here.

[#10]

Quote from: Tony

I've voted 'No' because I think that, on the whole, Plusnet have tried their hardest to cope with the spam situation.

Really? I don't call totally neglecting to train the spam filters for a week "trying their hardest"!

Quote from: pwebb

Quote from: rogerloxton This would lead to the assumption that DSPAM hasn't been trained in lieu of other projects that are ongoing. I have just checked and it was trained on the 24th and will be trained again today. The reason for this is not due to Postini or any other projects, but simply due to the reduced staffing levels over the Christmas period. Phil

[#11]

Quote from: Bob

People are talking a bit about 'hindsight'.

That would be me and I have found over many years that there is usually a certain amount of it involved in criticism of a process or procedure.

Quote from: Bob

There is an unfortunate habit in the 'computer' world to get the customers to test something.

That is something that is common in many areas...the motor industry springs to mind.
With respect to testing Postini or anything for that matter getting a test procedure that is close to real world conditions is usually extremely difficult if not impossible.
The varying nature of the real world and in the case of spam the constantly reviewing and adjusting of the source can make testing something of a best guess.
Most of us live in the real world not the ideal one and we have to live with it's faults

[#12]

I guess that it would have been better if, when people said they would be willing to join the test, that they would be selected and check how they had their email arrangements made. Then they could be asked not to change their set-up unless asked by PlusNet.
They could be given a positive avenue to report problems.
That would ensure that the trial was under control. Of course as has been mentioned sufficient staff should be to run the test. If you will pardon the expression of an old friend of mine,---- it was like 'old women's footprints'  -- all over the place. The trial was not under PlusNet's control.

[#13]

Quote from: Bob

I guess that it would have been better if, when people said they would be willing to join the test, that they would be selected and check how they had their email arrangements made.

Quite. Letting people on to the trial who had their spam setting on delete was bonkers!

[#14]

Quote from: jelv

Quite. Letting people on to the trial who had their spam setting on delete was bonkers!

No it wasn't - they did not suspect (any more than you did) that Postini headers would adversely affect DSPAM marking.
You only know that with hindsight. If you knew that from the start Jelv, why didn't you post to that effect in this thread and save all of those lost emails?

[#15]

The hindsight was more of the "d'oh - should have thought of that" than having dismissed it as a risk up front - I have never used the delete setting and so it had not crossed my mind. If it had I would have flagged it up.
I'm not paid to think about things like that - but people at Plusnet are. When staring the trial I presume there was a change control raised which should have addressed all the possible risks - no matter how small there was always a chance that DSPAM would throw a wobbly at the Postini headers and there could be false positives. I'm not sure even if that had been suggested I would have thought of the implications for those with the delete setting.