2017-04-20

I expected better

I had an interesting exchange with a customer this evening.

His internet has been down for an hour and he texted our "Major Service Outage" number saying so and saying he "expected better".

I'm not sure what to say to that - if we really have a customer that expects that a fault, which they have not even reported, is fixed within an hour, out of office hours... What does that mean? If we really have people with that expectation I am tempted to try and sell to them. I expect it will be many thousands of pounds a month for such a service, with multiple redundant leased lines, diversely routed, and a lot of 24/7 active monitoring and staff. I suspect we could do it, for the right price.

Of course, that is not what we normally sell. So if one of our customers expects that they have a serious misunderstanding as to what they have purchased. I try to be honest in what we sell, so I am worried that someone expected more.

At the end of the day things calmed down. I was unhappy as he "abused" our MSO system. It alerts many staff for what could be a major outage. In the evening, not so bad, but middle of night, not good at all. Either way it is staff that are not at work and trying to spend time with family. The reaction from staff that do get the alert was not good, and I took over and was, I have to say, rude...

As per the web pages on this, I was rude. You are asking for a rude reply if you abuse the MSO system! That is what we say!

It turns out he read of the MSO text details on the status pages, which does not have as many caveats. We need to fix that. It is clear it has to be "multiple lines" and so was clearly abusing the system. It was not as clear as the main web page on this. So we can be clearer, I conceded. We'll fix that. Sorry.

Ironically, this fault looks like it may affect 3 lines. It is even an "incident" that TT have spotted. So ironically not something we need to take any action on as it is all in hand. The issue is that this person did not know it was impacting more lines. Our aggregate systems do not spot three lines going down together as that is too few to recognise as a pattern. Good news is it should be fixed soon.

I hope the mix of rude and reconciliatory messages was the right level in the end. We'll see. At the end of the day, expecting a line issue to be fixed within an hour of *not* reporting a fault is special, in my opinion - and I'd be happy to sell such services for the price they deserve...

Let's hope this line is sorted soon.

26 comments:

  1. Out of interest, how often do you get a "valid" MSO message?

    Do all customers — even those with just one or two lines — need to have the MSO number, or is that just inviting problems?

    ReplyDelete
  2. Sorry - I mean the "MSO" keyword, rather than the number itself.

    ReplyDelete
    Replies
    1. It is getting rarer all the time as we, and carriers, get better.

      Delete
  3. On http://aa.net.uk/support-mso.html

    The text below the big yellow box

    "you'll get a rude reply and adding the to cried wolf list"

    should probably read

    "you'll get a rude reply and added to the cried wolf list"

    ReplyDelete
  4. Can't the MSO system ask users for at least two affected line numbers in the message and then validate them before alerting staff? If only one number, then reply "Sorry, not classed as MSO, but fault has been logged" (where you then start grouping by exchange/carrier to see if it is the first signs).

    Yes, it's some logic on the text message receiving side - but at least it'll stop/reduce false alerts.

    ReplyDelete
    Replies
    1. Obviously we could make it more complex, but I really don't feel expecting people to be able to read is that unreasonable.

      Delete
  5. Its odd the things that some customers expect. We don't use a queue on our support lines - we figure that our customer's time is precious so if all our support staff are on the phone (which is rare) they get dropped to a voicemail prompt and we'll call them back ASAP, rather than leaving them on hold. Usually they get a call back within 5-10 minutes and this works well.

    But every so often we get a customer with a problem that they consider really urgent, so they call repeatedly for a few minutes, dropping to voicemail each time but never actually leave a message. The customers who do this also usually withhold CLID so we wouldn't even know who it was unless they leave a message.

    Invariably, they give up and eventually call back a few hours later and complain about their "urgent" problem not being resolved because we never called them back (how could we?), whereas if they had phoned just once and actually left a message describing the problem, we would've probably fixed it for them within a couple of minutes!

    I never do understand some people...

    Oh yes, and don't get me started on people who report problems with a vague "not working" description and when asked for more detail (e.g. if they are getting an error message) just reply with "still not working"...

    ReplyDelete
  6. I was well aware of the 9-5 support before I joined. It's a bit of a downside, but I figured I'd rather have a competent 9-5 support service (And by and large it's been very good) than have you tried turning it off and on again between the hours of 8 and 10

    If you do have a outage at the weekend (Yet to be a problem) it does feel a little bit like there is nobody you can contact and nothing you can do (That said, the control panel is absolutely amazing in terms of capability for any line related issues).

    Perhaps some sort of aggregate single line outage text service would be a good idea. If there is a spike of reports, time to dig someone out of bed. At least that way it feels like you can report it, and if it's more than a single line fault then it'll get picked up.

    ReplyDelete
    Replies
    1. Bear in mind that their systems will probably notice an outage on your line before you do; pretty much every (rare) time I've gone to report an issue, they've already known about it (and, usually, fixed it!)

      My only real issue with the MSO number is why it even exists - if my line drops there's no easy way for me to tell if it's a major outage without diving onto IRC and seeing how many other folks are complaining. Your average punter isn't going to do that, so how could he possibly know it's a major outage?

      Perhaps replace it with a quorum vote on IRC? :)

      Delete
    2. People ask how they should know - the answer is that if you don't know, don't send an MSO. It really is simple. This is not really for a line outage anyway - if one line is down, not an MSO, and means typically engineer (working hours) so no need for an MSO system. If it is a lot of lines it will have been spotted by carrier or us and will be in hand. Where this has been useful is where there is some subtle routing issue over LINX or transit, or where we have some weird wide spread packet loss in a back haul carrier (something they do not usually spot), that sort of thing. That can usually be identified as a major outage by discussion on irc, or people that have visibility of many lines.

      Delete
    3. Oh, and of course, the unexpected - something we have never seen before, and so have no monitoring in place.

      Delete
    4. The times I've been on IRC during an MSO there has been a short debate whether there is one, then working out who is actually going to send it.. only for staff to turn up 30 seconds later saying their automated monitoring had already pinged them :p

      Delete
    5. To be honest, we don't mind that at all. The issue is people who know they can only see one line, their line, down, and down for a short period, and decide they will declare and MSO, especially if in the middle of the night. We had one person with an accounts query in the middle of the night decided it was an MSO!

      Delete
    6. A simple set of radio buttons, "how many lines are affected by this fault? 1 / 2-3 / 4+", with '1' giving the message "that's not an MSO, email (address)" should help I think. Or even stealthily divert single line "MSO" reports to the regular support queue...

      Delete
    7. Err, confused, how do we have "radio buttons" on someones phone when they send a text? And anyway we'll have people in the "how am I expected to know?" camp picking the highest number anyway on bases of "how am I expected to know?"

      Delete
    8. I meant the web page before getting to the MSO contact information - presumably such people go looking for "get my line fixed right now without waiting for office hours" and decide the MSO number is the best fit.

      Delete
    9. Oh yes RevK I remember being on IRC when the MSO was used for an accounts query. Couldn't believe it.

      Delete
  7. I think whats happening here is Customer becomes unhappy and/or feels entitled (for some reason) so issues an MSO to page staff, even in the night.

    ReplyDelete
  8. How about an automated reply to the initial MSO text asking customer to re-think, then only respond to confirm if they have a sound reason to believe that others are affected?
    Is this really such a frequent problem or is this just an isolated incident of the sort you're bound to get once in a blue moon, however clear you make things? Someone could be inebriated or high, or just in a foul mood at the time they lose connection. I would have though that 99% of A&A customers are fully aware of the purpose of an MSO text.

    ReplyDelete
    Replies
    1. It is not frequent - so far it is just frequent enough to allow us to add a blacklist but not to spend the time making a more complex system. When it happens it is annoying.

      Delete
  9. I love Adrian for this - he worries about the one in ten thousand (abusive) customer who is down for a period of time the average (admittedly non-AAISP) customer wouldn't notice.

    You're a good man.

    ReplyDelete
    Replies
    1. Thank you - I worry about anything that is not right, and work to make it right... Maybe I can get medication for this :-)

      Delete
    2. Probably a wee bit late for that :)

      You'll have to find other ways of being evil.

      /me points at BTOR :P

      Delete
  10. Interesting case this evening... 1 big blip (at 00:05) followed by 3-4 smaller ones. A few people noticed (not many due to the time, I guess) - Decided not to MSO since whatever was wrong seemed to be fixing itself (or the failover worked, same effect) and waking someone up probably wasn't going to help.

    OTOH not everyone that it happened to would know to jump on IRC - would it have been better to MSO so that status got updated or would that have just invited colourful language? :p

    ReplyDelete
    Replies
    1. Sounds like you did the right thing - maybe a normal text or email to support so we look in the morning. But, knowing it was multiple lines, we would have been happy for an MSO text. The issue is the people that see their own line down, don't know of any other, maybe done even do basic checks on their own line, but MSO for it. Or worse, expect us to have fixed this unreported fault in an hour and MSO.

      Delete

Comments are moderated purely to filter out obvious spam, but it means they may not show immediately.

Deliveries from China

I have PCBs made in China (well Hong Kong). This is all my many small PCB projects (not FireBrick). I would rather use UK suppliers but I am...