Interview with Matt / “Dockarl”

Posted on 22 August 2007 at 8:21 UTC, filed under People, disclaimer

Matt at Google Hi “Doc”, it’s cool to have you here! It’s great that the web removes barriers like the physical distance from here in Switzerland to Australia. Matt has been one of the regular contributors to the Google Webmaster Help Groups since January 2007. He has a diverse background: Agriculture and Computers, an interesting mixture, or how he puts it in his profile: “I know about cows and computers” :-) .

Looking at your first posts, I see a desperate webmaster, someone even screaming for “HELP!!!” in the thread titles. How did you find the Google Webmaster Help groups and what made you decide to originally post about your problems there?

Hmm.. how did I find the groups – I think I might have searched “How to contact Google” and came across the webmaster help groups there. I had to – I’d come across a problem that I just couldn’t get an answer to by doing a regular Google search, I knew it was an unusual problem and, like many other webmasters, I figured I might be able to find a real, living, breathing Googler somewhere to talk about the problem.

Did you get a satisfactory answer to your original questions in the groups? What elements were vital to that outcome?

Well, for some reason the answers to that post (it was back in 2006) have been ‘lost in the system’ but I did get a lot of hypotheticals from the regular group members – but nothing that helped, unfortunately.

How that came about is a very long story, but hell, you’ve asked, so I’ll tell you :-) . The person who owned the intellectual property we had been laboring to develop for the last two years had turned nasty – and was annoyed that we used their name on our website (and outranked them for it). My business partner and I were receiving ~20+ calls a day between us from the person. The phone calls started to elevate to the extent that we considered them threatening, and we were forced to call the police.

In the wash-up we just decided that – as a family business – we weren’t prepared to have to explain to my business partners kids (both under 5) why mum was crying and the police were ‘coming for a visit’ on a Saturday morning – so we decided to remove the name in question to stop further stress, even though we had every right to use it.

So I took the quickest path possible, made the changes to the website and asked Google to remove the cache. It had unintended consequences – it totally removed the ‘snippets’ from our website (our listings were title only), and we were left with a huge traffic decline. This, on top of everything else was absolutely crippling to the business. So, by the time I posted here I was getting a bit desperate – and it’s one reason I’m generally patient with people that come to the groups angry.

In the end, unfortunately no one here could give me the answer to the problem – it was out of their control. I hadn’t realized that a cache removal would remain in effect for 6 months. The main element that was vital to my outcome was Vanessa Fox (the beaut person that she is) who saw my post and stepped in and tweaked the system to let my site back in.

You’re a webmaster, you had issues with your site and Google and posted in the groups. If a webmaster came up to you and asked if it would be worthwhile to post about his problems there, what would you tell them? Would it make any difference if the webmaster was new to webmastering?

That’s an easy question. We’ve got a great community of beaut people here – you just don’t spend hours helping people gratis unless you’re passionate about it, so we tend to be universally ‘nice’ to people, especially newbies. I’d say ‘Go ahead, write your question, try to be succinct about it and TRY NOT TO PANIC!’. I’d also make sure that they knew that the people helping would more than likely be knowledgeable volunteers, so make sure you check your frustration at the door :)

What was it that made you stick around in the Google Groups, not only to ask more questions but also to help answer other people’s questions? What makes the Groups special compared to other forums?

Well I think that JLH and yourself made the effort to email me and help with some problems I was having with a hobby site of mine called ‘utheguru’ – that was an awesome gesture and made me feel at home. That kind of thing, along with the occasional guest post by a Googler, is what makes this forum special

In parallel to that, things had degenerated a lot further with our business to the extent that lawyers had become involved, and I had to put my PhD (and hence, income) on hold to spend my time dealing with that. I was looking for a stress release, and I’ve always been the kind of person that finds learning natural, cathartic and relaxing – so I got hooked.

If I’m honest, I also figured it was a way I could work towards another goal of mine – working with Google.

As an undergrad student, I read Page and Brin’s paper, and thought – “wow, that’s a neat idea”. The whole concept of Pagerank and linkages is something that’s really been around in science for hundreds of years. A good scientific paper is one that references other authors widely, and a reputable scientist is one that has papers referenced by many others. The CONCEPT of Pagerank is really nothing new in science – it just took a neat idea by those two fellows to convert the concept into something that could transcend academia and become relevant to that new thing called ‘the Internet’. Google became popular, first, amongst scientists – that’s something I observed and there was certainly alot of buzz about it within that sector of society before it ever became the household name it is now.

I’ve been a Google user ever since, and I’m fascinated by the system itself, how it works, the company, the culture – everything about Google appeals to me.

Further to the reasons Google fascinates me (you didn’t ask but I’m gonna tell you anyway.. haha), before the rather wild ride of backless lingerie began, I’d worked for some time as a Scientist with the Sugar industry (especially on the field / mechanisation side), and one of the major things I worked on there was reward algorithms – trying to use disparate manufacturing measures at the mill end of the system to send ‘quality’ signals to harvester operators. Hmmm.. how do I explain this – well, I’ve gotta go into a little background detail…

Sugarcane harvesters chop up cane into little lengths, about 8 inches long, called billets. Along with the cane, the leaf material is also chopped up. If that leaf material reaches the mill, it can have a bad affect on the quality of the sugar produced, and it also makes the cane more expensive to process and transport. So, the harvesting machines have big 6 foot metal fans which rotate at about 1000 rpm – that’s a phenomenal tip speed. These fans sit above the cane right after it’s been chopped, and their aim is to remove the leaf material. Unfortunately, a whole complex set of interactions conspire to result in a situation where if you try ‘too hard’ to remove the leaf material, you also end up losing about 20% of the cane you harvest through those fans – but it’s invisible. A billet that’s gone through an extractor fan ends up looking something like dessicated coconut – and there is no way of knowing the losses exist unless you do scientific trials to prove it.

I’d done the trials – all through North Queensland, in Papua New Guinea – all over the place. We had proved the losses existed, and the cost to the industry was in the billions of dollars per year, let alone the environmental impact. But because you can’t actually SEE the losses, you have a hard time convincing people that they actually exist. We got to the stage where my team and I had convinced the industry that there was a serious problem, and the next step was obviously “How do we stop it”. We knew that there was a ‘sweet spot’ where those losses could be reduced to around 5% depending upon the way the harvester was operated. Since we didn’t have the ability to measure what was happening in the field on a real time basis, we had no choice but to use indirect measurements in the mill – like fibre, the sweetness of the cane etc, to try and infer what was happening in the field – to measure ‘quality’ of the job.

That became my focus, and I learnt along the way that when you’re trying to make a reward system based upon derived measures, the tiniest little change to your algorithm can have huge impacts upon the system you’re trying to model. Also, if you’re offering “rewards” based upon indirect measurements, you actually end up becoming an intrinsic part of the system you’re trying to model – in clearer terms, the whole system tends to change or adapt to maximize “profits”, which can play havoc with the “accuracy” of your algorithm.

It sounds completely unrelated, but that’s actually Google (and the spam struggle) in a nutshell. That’s one of the reasons I’m fascinated with it and feel at home here in the groups where occasionally we get questions that make me think quite deeply about the challenges Google must face – and we get the opportunity to debate our views :) This thread about pagerank where Craig and I duked it out with full respect for each others opinion is one example I can think of that I’ve enjoyed.

You studied Agriculture and set up a shop to make and sell backless lingerie. I bet all the guys in the groups have visited your full site (for SEO reasons, I’m sure ;) ). How did that ever come about?

Ha – not only did I study Ag, but I managed to convince the government here to award me a scholarship to do a coursework Master’s degree in Computer and Comms engineering. I ended up with a few awards and an aggregate score of over 93% – without an undergrad engineering degree – I think that surprised everyone, even me :-) . But I guess it’s only natural – most people do best when they’re doing something they love. I’ve always been fascinated with those applications where IT, Engineering and Science intersect and meet ‘the real world’ – that’s kind of Googly.

An example – I can remember the time when I was about 12 years old that I blew up the family commodore 64 trying to get it to drive solenoids to water the garden for Mum. I didn’t realise at the time that you need a transistor and a relay if you want to drive something hefty like a solenoid with a TTL output :-)

But apart from being a bit of a terror, I’ve also always been a traveler and got along easily with folks. As such, when I was writing my Masters thesis, I figured I’d go stay with some mates overseas – I had a load of frequent flyer points I wanted to use, they all offered to put me up for free, so I figured it was an opportunity too good to miss. The only ‘gotcha’ was that I was to provide the beer – Norway was a hoot – my oh my – the Vikings ARE NOT dead!

I ended up (between parties) writing most of my Master’s degree tapping away on my laptop, perched on the edge of a fjord whilst staying with my Norwegian Marine Biologist friend in Northern Norway for a few months mid 2005 – the 24 hour sunlight was GREAT.

On the way back I dropped in to see my Indian mate in Tirupur (the south of India, in a state called Tamil Nadu) and ended up spending a few months there too. Tirupur is a big textile producing area, and I made friends with some of the big players there.

When I finally arrived back in Australia I mentioned that to my Brother in Law (a solicitor) and he said “well, I’ve got some clients that are looking to manufacture a neat new product they’ve developed” – so, before I knew it, I was off to India where I learnt all about ladies underwear, mobilon and thread density. We quickly got a few test shipments under our belt.

Upon returning my brother and I were asked if we’d like to get more deeply involved with the sale and promotion of the product – somehow I let myself be convinced. There began the roller coaster ride – I became manufacturer (traveled to China as well for that part several times), web developer, email wrangler, undy packer, book keeper, promoter and media spokesperson. It was crazy work and it was unpaid – the cost of manufacture and promotion sucked away much of my savings and any profit the product brought in before it ever had a chance to reach my pocket – although attending the modeling shoots was fun, and the POSSIBILITY that it might become something big was intoxicating!

But – a word from the wise – ever heard of Ali Baba and the 40 Thieves? Those folk were in the rag trade :-) Get involved at your peril.

One of your sites has recently had a strange kind of trouble with Google’s index, with all sorts of possible explanations but no resolution so far. For the average webmaster these kinds of situations are incomprehensible and terribly frustrating. What would you tell the webmaster when stuck in a rut like that – keep working on the problem or let it sit for a while?

First I’d ask them to think about whether they’d made any big changes to their site recently – to try and hone in on whether it might be something they’d caused themselves, rather than anything algorithmic.

Next, if I’d decided it might indeed be a penalty, I’d usually give them a copy of the webmaster guidelines and say “What do you think it might be?” – people usually have a fairly good idea about what they might have done wrong if a potential penalty is involved. I’d then ask them to write out a list of potential issues, and correct them + submit a reconsideration request and wait a month. If that didn’t work, time to put on the “mad scientist” hat and get methodical about things.

First I’d probably use Google to do a search for other people experiencing the problem. From there I’d approach these groups. If that drew blanks, I’d then start tweaking things with their site – but softly softly – one change at a time, waiting at least a week between changes so that I’d have a fair idea what ‘the cure’ was for future reference.

If that didn’t work I’d probably just start to assume that they were the victim of Google collateral damage – hell, we all know it happens, and I’d be submitting some attention grabbing posts to this group to try and ‘elevate it’ to the attention of Googlers, so that they could use their gadgetry to try and work out what the story was.

At that stage things are out of your hands, and you just hope that perhaps you’ve alerted Google to a potential “Googlebug” that might stop others from experiencing the same kinds of issues.

Assuming you had full access to Google’s servers and some web designers + programmers to help you, what would you change?

Hmmm.. looking back through my prep notes for my Google interview here…

I think I’d start with the problem of penalties. I’d be sitting down with the alg team and trying to thrash out a way that we could actually help those ‘ma and pa’ webmasters that have accidentally shot themselves in the foot – and to do so without giving the spammers a leg up.

I’d write out a list of things that we considered ‘top secret’ and another of those factors that were ‘out of the bag’, and I’d set about implementing changes to Google webmaster tools to alert folks to little things – like obviously hidden text – that might be resulting in a penalty and which they might not know about. Those kind of issues, to my mind anyway, are already well known amongst spammers and you can’t lose much by letting people know about them.

As for the more complex things, like, for example, keyword density (it’s a simple one, I know, but let’s start there) – you know, things that aren’t black or white – things where there were shades of grey, I’d be making tools to show them which side of the line they are tending towards – like a gauge, or traffic lights.

“We think your site is looking a little spammy – here’s an orange alert”.

Naturally, the alg team would then say to me “Well Matt, that’s all well and good, but if we start giving folks that kind of info, we’re essentially giving the spammers a great tool which they can use to test the limits of our alg, too”. I’d then say to them, well, why don’t we use cluster analysis to break sites down into 100 different categories of ‘spamminess’ – the traffic lights would just show how spammy you are relative to others in your ‘spamminess cluster’ – so really, if we give a green light to a known spammer, all we are telling him is that he’s kind of ok compared to the other spammers within his uber spammer group – but he needn’t know that :-)

For the spammers, the lights system would achieve nothing. For the ma’s and pa’s that are relatively innocuous, having a red light could be a huge help – just knowing you have a penalty lets you know that it’s actually something you can track down and correct.

But I suspect the other engineers would raise a whole load of reasons that my approach wouldn’t work – but I love the dynamics of a group, and part of the enjoyment of working in one is often the synergy that you find when you’re sitting down with a whole bunch of folks with common interests and intellect thrashing out a new idea – that’s how a lump of coal turns into a diamond.

That would be a plum position to be in.

After that I’d probably start gravitating towards the alg design / testing side of things – as that’s something I’m fascinated with – setting up mega test networks and conducting sensitivity analysis and pre-testing of new algorithm ideas would be lots of fun and extraordinarily satisfying – I love taking good ideas and helping make them better.

I’ve also thought I’d like to make a tool that shows a graphical representation of the linking structure of a site – with things like nofollow, noindex as an overlay – that could be a great troubleshooting tool for lots of problems too.

But, to be honest, most of my programming experience is at the nuts and bolts level – A GUI to me is a command line and a prompt – I’ve got a lot of engineer in me. I’d be able to write the crawlers and mangle the database, but I’d have to leave the bells and whistles to someone else :-)

You’ve done a lot of different things (so far, including an interview with Google). If you could rewind back to when you started studying, do you think you would do anything differently knowing what you know now (other than obviously buying some good stock)?

Cool! A rewind button!

Firstly, I wouldn’t have flown Qantas to my big interview – it was a debacle start to finish – they lost my bags (clothes, books, notes) my flights out (and back) were both delayed 12 hours or more and diverted because of tech probs – in short, I arrived sleep deprived and not feeling prepared, and I think I only hit my feet during the interview just after lunch. It was like an out-of-body experience.. grrr….

Secondly – I wouldn’t have studied Agriculture.

We had loads of fun out there, but my natural aptitudes are IT / Science / Engineering. My ag degree included a lot of that, but I tended to get let down by the sheer boredom of prac sessions that included watching grass grow – honestly.

I’m the kind of person that thrives on a challenge – so I did poorly at the “watching grass grow” practical subjects, and tended to dux the more academic subjects that others found a tad difficult – like advanced stats, biometry etc – I did the wrong degree for my skillset and, like it or not, time is a depreciating commodity.

I’m an extremely outdoors person, and I thought back then that if I studied IT or engineering I’d be stuck in front of a computer all day – but I now realize that that’s not really the case at all. Shucks, if I’m honest with myself, I LIKE spending time in front of the computer. I’ve come to realise that it’s the life / work balance that’s important – if you don’t have one, you tend to lose out on the other.

So with Ag, I just ended up naturally gravitating towards work that required me to be ‘stuck’ in front of a computer all day anyway, but getting paid poorly for it, so the opportunities to go outside and do adventurous things in your spare time were limited.

I’ve had some massive, great interesting experiences with the route I chose back then, most of which I don’t regret, but if I’d done IT or Eng instead of Ag, I think I’d be in a better place, career wise. You mention “good stock” – it’s funny that, because luckily I realized early that this wasn’t what I wanted to do long term, and tended to invest my wages well – so I’ve managed to have a decent lifestyle during the recent ‘challenges’ which is LUCKY :-)

Is there anything you’d like to add?

John Congrats on the new job, and I’m looking forward to achieving a dream like that myself soon, too – good on you mate! :-)

Thank you very much for your time and the replies, Matt!

There is one trackback ping to this post.
  1. [...] – those of you that know me well will know that there has been a stressful sidetrack regarding a new lingerie product over the last little while. I’m glad to say that this is now completely over and done with. [...]

There are 4 comments to this post.
  1. After graduating from the Vic20 I used the C64 to automate my HO model train switching, nothing as adventurous as an automated watering system.

    Nice interview Doc, now get back to blogging more so we can get those pages back in the index!

  2. I refer to my C64 experiments as “Peeks and Pokes and Magic Smoke” :-)

  3. I recently said to Craig I feel a bit cheeky hanging out with the big boys … but now I’m awestruck. I just cannot believe how generous you people are at sharing your wealth of knowledge and time with people like me. I enjoyed that, thanks.

  4. Very informative as well as entertaining!

Feel free to leave a reply to this posting.

Warning! Your comment will be lost if you mistype the spam-test or forget to enter your name or e-mail-address. Copy your comment to the clipboard to be sure.

You may use these tags within your reply: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>