The Special Business Unit

In which unqualified people inflict absolutely the dumbest ideas on us.

The story so far was that Edify had been bought by web banking competitor S1 Corporation, which in turn destroyed out web banking product in hopes of making our customers move to their platform, the major flaw in that plan being their product was garbage and their promised platform never seemed to materialize.  I think we changed 401k plan providers more often than S1 even did patch updates.  They were not good at the whole software thing.

The IVR side of the house was stuck into a building on San Thomas Expressway that is now gone, replaced by one of the big, swoopy NVidia buildings, declared the “special business unit” and left to fend for itself… but not before hanging a giant anchor around our neck in the form of the VP of sales, who was made CEO of the new unit.

Pretty sure we failed on all points…

I suppose the main lesson from this is never put the sales people in charge of a tech company.

The sales team is always somewhat at odds with the developers in an enterprise software organization.  They have quotas to hit and they get flustered because customers always want special features and they will promise anything to get them to sign on the bottom line.  And then they would show up asking for the feature they promised.  Sometimes it wasn’t so bad.  Sometimes it was insane.

It happened so often that back before the merger there was a limit set on a deal size before somebody could request a feature.  The potential deal had to be at least $6 million in size… and, suddenly, every deal that was being sized was potentially $6 million so the queue of feature requests, never slowed.  We used to joke that we should build up a price list based on all the features we didn’t have in the hope that the sales team would actually sell something we did have, their need to go off-menu being so reliable.

Anyway, they put Joe, a corpulent, red faced buffoon in charge of us and he set about creating a kingdom of his liking.  He wasn’t mean or nasty or anything.  He was quite friendly and jovial and people generally liked him as a person.  Joe was a nice guy in front of everybody.  He wanted to be liked.

But he was like the Moll Flanders husband who keeps living beyond his means while trying to get everybody on board to enjoy the party before it all comes crashing down.  And there were parties.  Many parties.  It was a hallmark of his era.  I don’t think he actually fled the company in the end by jumping out of a back window and riding off into the night, but I like to imagine that is how it went.  Anyway, Joe went straight to work screwing things up.

One of the first things he did for his sales team is start paying commissions when deals were put in the queue but not yet signed.  This led to a HUGE boost in unsigned deals in the backlog because the sales team was now incentivized to get customers there even if they had no hope in hell of actually closing a deal.  And that was easy enough to do.  You just tell the customer you want to get them in the queue but they don’t have to commit and they’ll go along.

That eventually had to be rolled back because it was patently dumb, so obviously so that I thought somebody was joking when they told me about it.  My first question was, “We’re paying commissions for not closing deals now?” and we were!

Our buffoon CEO also pushed our facilities guy… who was one of those people who signed up for every task offered until they were in way over their head and then commenced to make lame excuses as they failed to deliver on anything or meet any date… to sign a five year lease on our building because we were at the peak of the dotcom boom and rents were going up and the CEO, like sales people everywhere, thought the line would never stop going up so wanted to lock us in as projected rents might double or triple over five years given the behavior of the last couple.

Instead the dotcom bubble burst and as soon as a year later buildings in our complex were going for less than half of what we were paying and would continue to decline as 9/11 hit the economy.

Our CEO also wanted the company to become a leader in technology so signed a deal with an overseas university in a city that happened to correspond with where family lived so his trips were all business expenses, with first class airfare paid for by the company, to work on foundational research in speech technology.

This is something a giant, successful company can afford to do.  IBM did it back in the day, Lucent or Microsoft or Apple could manage it, and now big names like NVidia go there to keep their edge in the market.  A little company like Edify… we had no business doing that.

We also bought seats on the VXML Forum and SIP Forum standards committees… again, not something a little player like us should have ever done.  Lucent, Nuance, and SpeechWorks had bought in too hard before we showed up and the whole thing was about them squabbling of changes that would give one of them an advantage over the other two.

So we kept spending money even as sales were flattening out with the post-dotcom bubble reality.  We still brought in a lot of cash, but new sales were harder to obtain, so we were relying on our substantial ongoing service contracts to pay the bills.  Growth was not happening.

But that didn’t stop our CEO.  We totally missed our numbers?  Well, we’re still taking all the sales people on the big annual trip to Mexico, even if they didn’t make quota, because otherwise they might become demoralized and stop selling so much… as though their salaries didn’t depend on commissions.

We also had to hire the CEO’s daughter and best friend as summer interns where they basically hung out and gossiped and did as little work as possible.  This probably would have been the least of his sins if the pair hadn’t been so obviously there to party and do nothing.  Like, not even a pretense of work whenever I saw them.

And then there was the big annual holiday party, and the company birthday party, where he liked to stand up and bloviate about how great things were going and blah blah blah.

He also set us on a path to make a new product built on our platform.  This wasn’t necessarily a bad idea.  The web banking and HR apps we had done before had been very successful… so much so that they had led us to being acquired which got us into this situation with the buffoon CEO from hell.

The problem wasn’t building up an app, the problem was that it was envisioned as a “do all the things” contact center, customer relationship management app that was so ill defined that we spent a couple years working on it, the had to change it completely for the first customer.  It was a disaster that drew time and resources from the core product that was, and would always, make all the money for the business.

Meanwhile the CEO was printing up marketing documents based on interviews he did where he declared our platform could do things that we were clearly not capable of. There was a Dilbert cartoon at the time where the sales person promises that their product can detect tachyon emissions.  That was Joe in interviews.  And then he would loop back and make tachyon emission detection a critical priority.  So when he promised we had full language understanding in our product, he wanted it to happen.

We ended up buying a proto-language model startup called YY Software that was ahead of its time, but clearly not capable of delivering on its vision, and tried to make that into something.

That led to an incident where the team working on it was training the language model based on a history of email provided by the Royal Auto Club, which had been promised an automated email response system, and the lab’s test Microsoft Exchange server that was used to simulate the RAC environment and generate responses linked itself to the main corporate Exchange server, because that is what Exchange servers are supposed to do, and sent automated replies to the thousands of test email messages they had been working with.

That was not a good day.

There was also a huge push to get out features more quickly, so it was declared that we would go to a quarterly release cadence.

We had been on something like an annual release cadence, supplying patches and updates while we worked on the next release.  But sales needed features NOW if they were going to make quota… the poor sales people couldn’t be expected to stay engaged unless they had new shiny things to sell constantly… so we had the horrible year of four releases.

That may not sound like a big deal, but here is the thing.  IT departments do not like to upgrade software more than once a year.  Also, enterprise software vendors don’t like to support more than a couple of versions of their platform.  So everybody was happy with the annual cadence.  And then we sped it up and got a ton of push back from the field.

Our support policy, agreed to by all customers, was that they had to update to the current or previous version of the platform in order to get support.  By pushing four updates in a 12 month period, we basically pushed our entire customer base back four versions, so technically nobody was supported except for those few companies who happened to be ready for an update or were doing new installs.

And when push came to shove, when support started telling people about the policy, the threats to cancel the support contracts started to show up, and that was the life blood of the company, what paid the bills.  So the CEO… and the new VP of sales, who was a pretentious, self-important jerk of the “this feature should be easy, development is lying when they say how long they need to make it” sort… declared that we would support many versions of the platform, leading to no end of headaches in engineering because every release had to maintain its own build system because annual releases meant we often made big changes to libraries, third party integrations, and compiler versions.

We ended up having one version released during that cycle, EWF 7.2 as I recall, used by exactly one customer, Toys R Us.  That is not the situation you want to be in.  We also had one of those releases that had zero customers adopt it… which was better, but we had to go through all the supporting work for a release, which is not trivial, and we incremented the latest supported release by one with this.

And then there was the overwrought customer feedback module.  Joe gave some academic who had impressed his with something about “delighted” customers being some huge percentage more lucrative to a company, so Joe hired him at some no doubt obscene salary, put him in charge of this new product module, and told us to do what he said.  Again, a nice guy, but academics in a business setting are often completely out of their depth, especially when trying to shoehorn their pet theory into product reality.  He had absolutely no idea what would sell, he thought just having a good idea was enough.

So we ended up shipping a post-call survey module that was supposed to come with a bunch of professional service in the form of this guy setting up the surveys and then evaluating the data.  Because of this he insisted that every copy of the module come with a fully licensed copy of SPSS.  That is a bit like requiring the full version of Microsoft Access Professional with a simple contact manager app… only, at least Access is aimed at business use, and not a purely academic tool like SPSS.

And, to top it all off, I was told to sign off on the product before it was anywhere close to being done or ready to ship because it has been promised for whatever quarter that was.  I did what I was told, though I was comforted by the fact that I knew we would never sell a copy or, if we did, we would probably have to refund it for being a ridiculous waste of time.  Our professional services team had been making post-call surveys for years.  It was not a difficult thing.  Most companies just want a feel for how things went, not an academically rigorous study of their customers mood.

This was an era of lying really.  I am going to sound mean here, but that is what you get when you put the sales VP in charge of things.  This was the era when I was asked by marketing to create a list of features we were working on for the Gartner Magic Quadrant review and I gave them the true and honest list.  They came back and told me we were currently telling people we had all of these features already.  I had to ask what they wanted then.  They worked in marketing, they could theoretically make up better ideas for Gartner than I could.

With all of this profligate spending, the dotcom bubble having burst, our rent being way over market, and the crazy, go nowhere projects that were being demanded, and just a constant stream of lies to the market, our customers, our owners, and anybody who dared give us a listen, we were losing money.  That was made up for in the usual way, by laying people off.

Come 2010 and the valley crisis caused by the collapse of the housing market investment scam HR professionals started a policy of not hiring people who had been laid off, as there was something wrong with them.  There was a headline in the Wall Street Journal in favor of this.

I always felt you ought to get credit against that asinine policy for the number of layoffs you managed to survive.  I would have gotten a lot of credits as for several years we were doing layoffs every six months as we failed to control spending but needed to cut costs.

By that point I was a manager and had to stack rank my team every time that came around, prioritizing the people I would need to complete current projects… because at no time were we allowed to reduce our goals as we laid people off… so I had to pick and choose from who could cover what aspects of the plan.  Anybody who specialized in one area was likely gone… unless nobody else knew that code.  Then I couldn’t part with them.

That didn’t stop the parties.  We laid off a bunch of people one week and had a big holiday party at a fancy restaurant the next.  Or maybe that was the one at the museum of modern art?  They were always poorly planned by Joe’s admin, the hallmark of their incompetence being ritzy locations with about half as many chairs as people attending so early arrivals would grab all the seats and anybody after had to eat standing up.  Always the same mistake, over and over.  It was as though our corporate motto was something about never learning from our mistakes.

Finally, at one point, many quarters too late, Joe was called on the carpet for not delivering on anything he had promised and losing money… because all S1 asked of us was to not be a drag on the bottom line until they could sell us.  Just don’t suck was literally the goal, and he couldn’t manage it.

He was out as were a few other people at the top.  There was also another round of layoffs.  And then we got a new sheriff in town, a guy name Mitch.  Somebody who wasn’t in sales, who was supposed to have some business sense, somebody who was going to turn the place around.  A neighbor and pal of Steve Jobs who was going to do for us what he did for Next… quite literally in the end.

Mitch showed up on the day after another big layoff and told those of us who remained how we were going to return to growth.  He told us a bit about himself… he liked to remind everybody he was friends with Steve Jobs… or at least a neighbor of his… then ended up in a meandering, old man tale about his career where he told us he had never been dumb enough to hang around at a company when it was clearly going down hill and laying people off.

Then, of course, he realized that we was talking to a group that would be critical to any success the company would have going forward and that he had basically called us a bunch of losers and dumbshits for sticking around.  Also, it was not lost on us that mister smarty pants had also signed on for this voyage.  Anyway, he tried to recover in a way along the lines of that scene in Caddyshack.

This is the worst hat I have ever seen!  Does this hat come with a free bowl of soup?

Then on seeing the whole room wearing that hat…

But on you it looks good!

So we were leery.  He said a lot of the right things, but he clearly had a low opinion of us.  And, as would be obvious in the end, he didn’t give two shits about anybody in the room beside himself.  He didn’t care about the company at all.  He was just there to sell us down the river and collect his 30 pieces of silver like any of the VC/Wall Street class.

But first he wanted to get us out of our crappy building.  At least we were all kind of on board with that.  That led to the event which I refer to as “the move.”

The tales so far:

• Hanging on the Telephone
• Calling the Mall
• Finding Friends with the Popcorn Lady
• Jenny and the FCC
• What Your Area Code Used to Say About Your Location
• I Get a Modem then Start a BBS
• Tech Support at the Village
• Dominating the PowerBook Modem Market
• The Death of eWorld
• The Great Dial Tone Drought of 96
• When the Modem Business Ran Out at the Village
• The Village Attempts to Move Beyond Modems
• Anecdotes from the Village
• Usenet Newsgroups Part I – I find some Usenet Archives on CD-ROM
• Usenet Newsgroups Part II – What is on that CD and Why Your Hard Drive isn’t a Database
• Usenet Newsgroups Part III – Founding, Fame, Influence, and Foreshadowing
• From the Village to the Island
• YoYo Mechanic
• The Great Failure of Residential ISDN in the US Market
• Set Adrift from the Island
• Thinking on Enterprise Software
• Stock Options and the IPO
• Into the Telephony Subsystem
• Rambling about Speech Recognition
• Web Banking, The Acquisition, and the Start of the Great Decline