Opinion: GetAhead 101 is no way to manage networks

My advice to CXOs and network managers is this: hunt these deranged fools down like dogs

There it was. On page 143 of the new book by Nassim Nicholas Taleb, The Black Swan. In one sentence, the explanation for that modern, corporate obsession with appearing busy.

“When you are employed, hence dependent on other people’s judgment, looking busy can help you claim responsibility for the results in a random environment.” Taleb goes on: “The appearance of busyness reinforces the perception of causality, of the link between results and one’s role in them.”

My experience of working in corporates could not be more accurately described. The manipulation of the perception of connection between effort and results (and then making sure everyone knows about the “right” ones) is GetAhead 101.

Arthur Anderson, the now defunct accountancy firm, used to publish a league table in its office in London that represented how long each staff member had been in the office each week. The implication was that those who hung around in the office the longest were the most productive. Can you imagine how such a ludicrous idea could gain a foothold amongst the brains at a blue chip brand like AA? I think Taleb has the answer. If you’re a mediocre middle or senior manager, what better way of stacking the deck in your favour?

Now, you may wonder why I’m raising this in an article that is supposed to have a telco focus. Well, there are two reasons. The first is that I simply couldn’t resist sharing that excoriating insight. The second is that, as Epitiro’s broadband customer experience monitoring (CEM) role expands, this whole notion of cause and effect, and making an accurate connection between them, is becoming increasingly dear to my heart.

As you can imagine, as more clients get sight of our data, more theories abound as to why customer experience fluctuates, and (ultimately) whose fault it is. What’s interesting for me in all this is the fact that there is no existing historical data set to obtain guidance from. This data is unique, but more significantly, it’s new. Consequently, a) pretty much anyone can have a theory on why the data says what it does; b) it can be hard to prove them wrong, at least in the short term; c) the number of factors we find to be causes grows almost exponentially; d) it can take a while to get the hang of reading the data properly.

I have yet to hear that the reason for slow HTTP speeds is due to the reflection of squid boat lights off flocks of mutton-birds, but I’ve heard some theories that come pretty close. Frustratingly, it can also take some time to de-bunk these wilder theories, because the lack of historical data means we don’t have months of results that show that the world isn’t actually flat.

The guys I enjoy working with the most on the data, are the grey-haired network engineers. Their approach — because they’ve learned the hard way — is empirical. That is, they have to spend quite a lot of time making inductions from observed instances, or to you and I, living in the real world. The hardest people to work with — and, unbelievably, some of them are network engineers — are those that think along the lines of; “Don’t bother me with the facts, my mind is made up.”

It’s the ISPs these guys are over-represented in that are at the greatest risk of failing to improve their customer experience. I hope Epitiro’s customers would agree that at least some of the data we have provided them with has assisted the more empirically minded network specialists in identifying and fixing some network flaws.

The performance of one of our larger clients has improved directly and significantly as a result of it identifying the cause of its hitherto bottom of the industry DNS performance. It invested in a caching solution that was implemented specifically to address the cause. It then progressively turned it on. Our data showed corresponding decreases in response times, as the system was increasingly activated.

However, there are guys out there, clutching GetAhead101 tightly to their chests, who deny that there is anything anyone else can tell them about “their network”. Let’s de-construct a phrase I have heard more times than there are brain-cells in the universe:

“Oh yeah, we have a few probes in the network, they run 1,000 tests every second on most of the key performance indicators. We then review the results and make assessments as to what they mean, and act accordingly”.

For “Oh yeah” read: “Of course I’ve thought of it.”

“We have a few probes..” read: “We don’t really need anything to tell us what is going on in the network, but I looked busy and burned a fair bit of budget building these things so I have deployed them to show I wasn’t wasting time or money.”

“In the network” read: “As far away from the customer as possible.”

“They run 1,000 tests every second” read: “We’re generating a lot of activity, (just don’t ask me about the results).”

“On most of the Key Performance Indicators” read: “On a couple of the things we know we can measure.”

“We review the results and make assessments” read: “We bank the good and bin the bad”.

“And act accordingly” read: “We talk fast and throw in lots of acronyms so no one understands what we are talking about and subsequently get left alone so we can carry on looking busy.”

To you CXOs and network managers out there reading this, my advice is this; hunt these deranged fools down like dogs. Well, er, within the requirements of the Employment Relations Act of course. Or at least recruit empiricists (that is, those trained to search for the truth), not priests (those who believe they own the truth). Unfortunately, I suspect there is a direct correlation between brains and the former, and the lack thereof and the latter. Some Compuerworld readers will kill me, but I reckon the secret to being a good techie is to be interested in what you don’t know and be upfront about it; rather than presenting what you believe to be the case as irrefutable fact.

I’ve been mightily impressed by the way our existing ISP clients in New Zealand have taken our data, and attempted to prove it wrong at every turn. When they can’t, they know they have a problem. They find it, fix it and their performance improves.

Our local team spends a significant amount of its time working with these guys to understand the data, because to be honest, much of it is new to us too. As VoIP and video services roll out, this data will become mission critical, because customers won’t accept a VoIP service that delivers less than POTs, or a video service that cannot compare to TV picture quality or reliability.

The successful ISPs of the future will be those led by managers who understand the potency of empiricism, and can spot it — or the lack of it — in their teams. Excessive activity is the smokescreen of the mediocre.

Perhaps for the first time, customer experience monitoring will provide senior managers with the tools to spot those in their teams engaged in generating activity, and those engaged in generating results. Their customers will certainly be feeling the difference.

Cranna is managing director of broadband benchmarking company Epitiro Technologies in Australasia

Join the newsletter!

Error: Please check your email address.

Tags Networking & Telecomms ID

Show Comments
[]