Late last night, in the middle of an email about two optical transceivers, I wrote: “I bought like 2million in hardware today.”

The small order was for a pair of Juniper-compatible 10-gigabit modules headed to our New Jersey data center. Somebody else had found the same parts after I had already ordered them and asked whether to cancel theirs. My answer was to keep them. If we ended up with extras, we could return them.

The sentence about two million dollars was not a polished announcement. It was a tired message from my phone after a long day of getting year-end infrastructure purchases over the line.

That is what makes it useful.

Four days earlier, I had laid out the larger Juniper network proposal. We had negotiated the cost far below list, but the pricing depended on completing the deal before the calendar turned. By December 31, the approvals and orders were moving, cables and optics were being checked, and all the tiny dependencies around a large hardware plan were showing up at once.

Big purchases are made of small parts. The network can have an approved multi-million-dollar design and still wait on the correct transceiver. A spreadsheet can reconcile perfectly while the loading dock, rack space, power, shipping address, or delivery sequence remains wrong.

There was a lot happening around us. We were preparing the Empire State Building office, planning more work in New Jersey, and continuing to add capacity for a public company that was growing quickly. Hardware did not arrive as an abstract capital expense. It arrived in boxes that somebody had to receive, inventory, rack, cable, configure, monitor, and eventually replace.

The scale was exciting and a little absurd. Earlier in my career I worried about individual drives and small server orders. Now one day could carry roughly two million dollars in purchases, and the immediate conversation could still be whether we accidentally had two extra optics.

That feels like an honest final note for 2013: huge growth, very real stakes, and the work still coming down to the details.

Archive