Search This Blog

Thursday, May 29, 2014

Wi-Fiction

Here's another estimate on the cost of upgrading schools' network infrastructures to give every student Wi-Fi access.  You may recall I estimated $5.5 billion.  Then Funds for Learning estimated $10.8 billion.  Now CoSN (the Consortium for School Networking) and Education SuperHighway (ESH) have estimated $3.2 billion.  The differences are explained by the methodology.  My estimate was just extrapolated from a single data point.  Funds for Learning created a per-student average based on what schools had been requested, and applied that to all schools.  CoSN/ESH created a per-classroom basket of goods to calculate a per-school cost, and added district-level gear.  It then multiplied those costs by the number of schools and districts in the U.S., and came up with $2.9 billion.  And then added $300 million for libraries based on...nothing.

Who's right?  Nobody.  Obviously, my estimate is unreliable, since every school is different, and I used only one district with 2 schools.  FFL's estimate is bound to be too high, since it's based on what high-discount schools spend, and we know they overspend because 90% is too close to free.

CoSN/ESH's estimate suffers from being based on a theoretical configuration and average prices.  The quantities are reasonable, but I can't say they're right. I find the prices to be very optimistic.  $520 for an access point, installation included?  I'm seeing hardware prices around $1,000 for a manageable 802.11ac access point.  And what switch offers 6 gigabit PoE ports for $73 (OK, you don't need PoE on all ports, but you sure do for those access points)?  $12/port installed?  For a managed gigabit PoE switch, I'm seeing closer to $20/port for the hardware only.

CoSN/ESH also tried to estimate what percentage of schools actually need the upgrade.  That's a good idea, but I don't like their estimates for two reasons.  First, I think their view of current infrastructure is too optimistic.  For instance, they say that 43% of school LANs are ready for 1:1.  That's based on a CoSN survey of 472 tech directors.  Are those 472 representative of the nation?  I'd argue that they're likely to be more forward-looking since they are CoSN members, and proud of their networks since they volunteered to answer the survey.  Still, it's probably the best data out there.  But what is the actual data?  The question was, "How confident are you that the typical school's wireless network would have the capacity to handle a 1:1 deployment this fall?"  Only 15.2% were "very confident";  26.6% were "somewhat confident."  So what the data actually says is that 41.8% of tech directors surveyed are at least somewhat confident that most of their schools are ready for 1:1. It does not say that 43% of school networks are ready for 1:1.  It gets worse: in CoSN's report, we learn "no district reports full access in more than 71% of its schools."  And that's any wireless access, not enough capacity to handle 30 devices per room.  So 0% of the districts surveyed are completely ready for a 1:1 initiative.  I think all their estimates are overoptimistic.

But the bigger problem is that they are assuming that schools which have adequate infrastructure installed will not make upgrades.  If you offer someone an 80% discount, the meaning of "necessary" changes.
  • If I have to pay full price, I can get by with my current 802.11n-AP-in-every-third-room network.  But if I can get an 802.11ac AP in every room at 80% off, then that's what I'll do.
  • Is it necessary to upgrade that 5-year-old switch to a new managed PoE switch at a cost of $1,200?  Maybe not; we'll have to run an extension cord to the access points and suffer with 100 Mbps ports.  If the new switch cost $240?  Oh, yes.  
  • CoSN/ESH say 80% of schools have 6 drops/classroom.  Maybe.  Do they have 6 Cat6 drops ideally placed for future needs?  No, we installed Cat5 (which can handle gigabit on a good day), and hadn't anticipated the need for drops for an access point in every classroom, not to mention one for the smartboard and one for the projector.  I can't afford the $200 for a cable run to replace that 40' patch cable snaking up the wall and across the ceiling to the projector, but if I could get it for $40, then yes.
  • Late 2015, we'll see Wave 2 (MU MIMO) 802.11ac APs.  Time to replace all the APs.  And the upcoming 802.11ad (WiGig) sounds perfect for 1:1.  Each WiGig AP will deliver multiple gigs of data, so we need Cat6 cabling and 10 Gig PoE switches. Time to replace everything!  1-in-5 Rule?  Guess I'd better upgrade to Cat6 and 10 Gig PoE everywhere now, so that I'll be ready for WiGig.
If Priority Two funding becomes available to 70% applicants, look for a tsunami of forklift upgrades.  Even for 40% applicants, the demand will spike.

So I find the $3.2 billion figure to be a significant underestimation.  I'm liking my $5.5 billion guess more and more.

My final criticism: the report says that we'll be able to spread the $3.2 billion across 4 years.  Really, who's going to wait 4 years?  I suppose there are some schools who are upgrading this year who won't want to upgrade for 4 years.  But I don't think we can count on a quarter of applicants waiting 4 years.  Wireless access points get better every year, and they're easy to replace.

My crystal ball says requests for Wi-Fi gear will be: $4.2 billion in FY 2015-2016, $2.3 billion in FY 2015-2016, and $1.8 billion every year after that.  If maintenance is eligible, be ready for $500 million per year.

When exactly will the Chairman be convinced that $2.4 billion per year is not enough, even if he plays some game with the accounting to give the fund a one-time $2 billion goose?

Wednesday, May 28, 2014

Cram that funding!

I like libraries.  And librarians.  But the recent letter from the Urban Libraries Council to the FCC is way off base.  I agree with their basic statement, that libraries are funded less than school districts, but then our thoughts diverge.

Proportionality: The ULC makes the odd assumption that E-Rate funding should be apportioned based on the number of buildings.  I agree that on average, libraries get less funding per building than public schools. But I completely disagree that E-Rate funding should be allocated on a per-building basis.  And even if I agreed, their calculations are off.

The ULC compares the number of library buildings that exist to the number of school buildings included on E-Rate applications, and wants to give libraries funding based on that proportion.  Let's fix those numbers.  First, it's 16,392 library buildings.  Second, it's 111,517 school buildings on E-Rate applications.  Third, if you're going to include all libraries, whether they apply or not, let's include the 26,598 private schools that didn't apply (that number's probably a little low, since it counts schools, not locations, but it's good enough).  I don't have the number of public schools that didn't apply, but lets assume that the 976 school districts that didn't apply had an average of 1.02459 schools, so that we get a nice round 1,000 schools (hey, if I'm going to assume, why not give myself a round number).  So the total number of school buildings is 139,115.  Wait, the NCEF says 132,270, so let's take their number.  The ULC said that libraries should receive one sixth of all funding, but the truth is they only have about a ninth of the buildings eligible for E-Rate funding.

Then the ULC says that libraries only received $60-70 million out of $3.96 billion.  OK, first of all, the $3.96 billion is a fantasy based on what the fund would have been if it had been indexed for inflation.  How much do libraries get?  I only have figures for FY 2011-2012, so let's take a look: $71,796,246.10 in disbursements.  OK, the ULC did say "about $60-70 million," but I'm getting a little tired of how they always seem to rounding in their favor.  But that's disbursements.  We should be comparing approvals to the total size of the fund.  Even better, let's compare approved funding for libraries to total approved funding.  I'll stick with 2011.  Libraries were approved for $92,021,445.61 out of a total of  $3,116,629,832.23, or about 3%.  If libraries had gotten a proportion based on the number of buildings eligible for funding, it would have been closer to  $343,650,672.06.  Not the $560 million the ULC lays claim to.

Now let's look at what ULC claims results from this funding injustice:
  1. very few have 1 Gbps bandwidth to the building
    That's a cause, not an effect.  If more libraries chose to have 1 Gbps bandwidth, they would eat up a larger proportion of the fund.  If libraries had asked for E-Rate funding for 10 Gbps per building, they would have gotten it.
  2. perhaps none have the minimally adequate 5 Mbps downlink Wi-Fi per user at critical times
    "Perhaps none"?  I could say, "Perhaps all have 50 Mbps per user"; we don't know what bandwidth they have.  Also, 5 Mbps per user is not "minimally adequate"; well-equipped schools put 20 devices on a single access point with 60 Mbps throughput (802.11n may have 100 Mbps in bandwidth, but not in throughput).  That's 3 Mbps at all times, critical or not.
  3. few have adequate desktop computers for their user base
    Again, that's a cause, not an effect.  The E-Rate does not fund computers.  I agree libraries should have more computers.  And more books.  And more staff.  And more space.  But the E-Rate doesn't fund any of that.
  4. only a very few can afford the high cost of digital information
    I assume by this they're talking about online databases like Lexis-Nexis or JSTOR, which charge fees to access their information.  Again, not eligible for E-Rate funding.
No argument with the rest of this section: libraries are crucial resources for Internet access for many Americans.  It's a very good thing that 100% of libraries offer Internet access to the public.

Needs:  Public libraries receive less financial support from the federal government than any other institution in the civic landscape.  Really?  Let's see what the ALA says: LSTA: $180,909,000; grants to states:  $154,848,000; National Leadership grant: $12,200,000; Laura Bush 21st-century Librarian grants: $10 million; Native American and Hawaiian Library Services: $3,861,000.  That's a total of over $361 million.  Add in $92 million from the E-Rate, and we're at $453 million.  Meanwhile, museums get around $150 million/year (That doesn't include the Smithsonian, but the library figures don't include the Library of Congress).  And I just picked museums off the top of my head.

The undersigned believe in a two-part formula:: (1) income of the user group (weighted by cost of living), plus (2) number of daily users of the building (because the number leads to assessing the necessary Wi-Fi and desktop connectivity).  I'm OK with number one; weighting cost of living adds a new layer of complexity to the discount matrix, which is bad, but it will drive up demand for E-Rate consultants, which is good.  But number two is wrong.  The need for connectivity need is driven by simultaneous users, not daily users.  And by "user," I assume the ULC means someone who uses the library's Internet connection, not someone who visits the library.  Because I probably use my library 100 times a year, but I never use their Internet access.

Because a large urban or suburban library will have at least as many users per day as there are students in a large high school.  Let's see what kind of data we can find.  On page 36 of this report, the IMLS says that library buildings serving a population of 1,000,000 or more have an average of 26.8 public computers.  I can't find any broad statistical studies on Wi-Fi usage, but the the Buffalo Public Library (not a ULC member, but a large urban library) publishes its stats online.  It's main branch seems to have about 5,000 logins/month. That branch is open 6 days/week, roughly 25 days/month, which means about 20 logins/day.  Let's say that all 26.8 public access computers and all 20 Wi-Fi logins use the computers at once.  That's 46.8 simultaneous users.  I'll round it up to 50 simultaneous users.  In the average urban school, there are 3.4 students/computer.  I checked the Buffalo Public Schools high school enrollments, and the largest high school I could find had 1,080 students.  That doesn't strike me as a large high school, but let's go with it.  Assuming Buffalo has a typical student/computer ratio, that's about 317 computers.  So does that urban library with a max of 50 users at a time need as much funding as a school with 300 computers?  And the other library branches in Buffalo have 10-25% of the volume of computer use of the main branch.  Do the small branches deserve as much funding?

Administration: No checkable facts here.  They suggest 3 ways to lower prices for libraries:
  1. All libraries should have access to the contracting prices obtained by other libraries and by schools in similar geographic areas.
    Hear! Hear!  Give us Lowest Corresponding Price.  Or at least publish price and product information hidden in the Item 21 Attachments.  I know I'm dreaming, but at least we have heard the Chairman suggest that the FCC should publish pricing benchmarks.
  2. All public libraries should be able to opt into contracts that the FCC itself puts out for bids.
    The FCC probably doesn't bid out any contracts.  That's handled by the GSA.  And honestly, a contract to serve the FCC is probably not going to be of much use to libraries outside the DC metropolitan area.
  3. All public libraries should be able to know that they can contract for 'whole networks.' This means access to the Internet at a wide area network point of presence, a 1 Gbps fiber connection to every library building (two thirds of libraries have no fiber and those that do cannot afford the electronics upgrade to Gbps bandwidth), a 5 Mbps Wi-Fi downlink inside all buildings, as well as caching, firewall, and maintenance.
    A laudable goal.  But we're going to need a bigger fund.  
So let's look at the real reasons libraries don't get more E-Rate funding.

First, a lot of libraries don't apply.  Here are the reasons why.  The top 3 reasons:
  1. CIPA.  Those dang librarians just won't knuckle under and sacrifice our civil rights to get their hands on some federal moolah.
  2. "The E-rate application process is too complicated."
  3. "Our total E-rate discount is fairly low and not worth the time needed to participate in the program."
Second, libraries get less funding because their telecom and Internet needs are lower.  If libraries requested more funding, they would get it.  Let's take the first signatory on the list whose service area is a single school district: Alexandria, VA.  We'll look at pre-discount telecom and Internet costs for FY 2014-2015.  Library: $35,773.68.  School district: $791,172.00.  Let's make that a per-building cost. Library: $8,943.42. School district: $ 34,398.78.

And no, it's not that libraries are falling further short of their needs.  More than half of libraries said their bandwidth is sufficient to meet patron needs almost all of the time.  When libraries rated the challenges in providing access to patrons, the top  problems were: quantity of staff, staff training, insufficient workstations, liability issues, and the library's time limits on access.  Bandwidth came in 6th.  According to one survey, 37% of schools have inadequate bandwidth, compared with 34% of libraries.

If the ULC wants to increase the proportion of E-Rate funding going to libraries, they should campaign for changes in CIPA regulations and simplification of the application process.  The solution is not to give libraries more funding than they're asking for.

Monday, May 26, 2014

Cost-allocating the costless

The FCC is closing the barn door just when it seems likely they're going to get rid of the horses.  On Friday, the FCC released what may become known as the Bundling Order.  The order says, in a nutshell, "beginning with funding year 2015, E-rate recipients must cost allocate non-ancillary ineligible components that are bundled with eligible products or services...."

What's that you say, that was already the rule?  Well, yes, but since the beginning of the program, USAC and the FCC have turned a blind eye to free cell phones.  (In addition to turning a blind eye to use at ineligible locations and the fact that the IRS long considered employer-provided cell phones to be a form of employee compensation.)

But when USAC read the Sixth Report & Order, they made it clear that free cell phones were not allowed.  Cell phone companies squawked, so the FCC released the incredibly misnamed Clarification Order, which tried to create a loophole to allow applicants to get free cell phones.  Then some cell phone companies tried to expand the loophole by offering free tablets, and some VoIP providers offered free desk phones.  The gold rush was on.  In July 2012, SECA asked the FCC to clarify the Clarification Order,  and the FCC quickly released a Request for Comment.  The muddle went on, much to the chagrin of responsible service providers.  And now 20 short months later, we have the order.

The order is good. First, it takes effect over 13 months from now, so applicants have time to get their cells in order.  More importantly, it changes the rule to what I said it should be: bundled equipment should only be allowed if it's ancillary.  (OK, my first suggestion was that the FCC should not be in the business of regulating the procurement practices of local governments, but since it's clear that they can't keep their noses out of procurement, I had a fall-back.)  And no, cell phone handsets do not meet the requirements for Ancillary Use.

The order is pretty clear, but there are a few gray areas.

The big question: How exactly do you calculate the cost-allocation?  It seems reasonable to deduct the full cost of the phone (which is not hard to find, just go on your carrier's website, pick out a phone, and get the "no-contract" cost for that phone) when the phone is received.  Two problems with that: 1) I don't think USAC's forms will allow you to enter a negative one-time ineligible cost, especially if the total one-time cost is zero, and 2) in the case of plans with shared minutes, the cost of 12 months of service on a line is often less than the cost of a phone; if you start such a plan with, say, 10 new phones and a shared pool of 700 minutes/month, the total pre-discount price will be a negative number.  (If your funding request is negative, do you have to pay money into the fund?)

It seems like USAC or the FCC should clear this one up.  Fortunately, they have many months to do it before applicants start the application process for 2015-2016.  My suggestion: have applicants take the cost of the phone, divide it by the number of months in the contract, and enter that number in Item 23B (ineligible monthly costs).

But the math gets worse: What do you do when you have a smartphone?  Now you have a telecom FRN and an Internet FRN, so do you just split the cost of the phone across the FRNs?  But wait, off-campus Internet access is not eligible, so applicants would have to cost-allocate at least part of the cost of the Internet access.  So let's say voice service costs $40/month, while a cheap data plan costs $15/month.  If you split the $650 cost of an iPhone 5s, each FRN gets a $325 charge.  Divide that by 24 months, and you get a per-month cost of $13.54 for each FRN.  Now let's say that the intrepid school employee who uses this phone spends half her waking hours on school property, even on weekends.  So half the data plan is eligible (while it's true that the phone is receiving emails while the employee is sleeping, it seems harsh to count those hours in determining percentage of use at eligible locations).  So the data FRN would have $15 total monthly cost, minus $13.54 for the handset, minus $7.50 for the use at ineligible locations, for a total of negative $6.04.  That scenario is not likely for most individual plans, since $15 is a really small data plan, but it could well happen with a shared plan (for instance, on Verizon, 25 employees sharing 50 GB would average out to $15 a phone).

Having applicants create a per-month cost for the handset creates another question: What are applicants in the middle of a multi-year contract supposed to do?  In general, you only get a free phone with a 2-year contract, so come July 1, 2015, about half of applicants will be in the middle of a contract.  You could make a case that the gift (free phone) was given back when the loophole was allowed, so applicants shouldn't have to cost-allocate in subsequent years.  But we pointed out above that the cost of the phone should be spread across the term of the contract.  And it's clear that the carrier is recovering the cost of the phone over the course of the contract (witness the large early termination fees).  It's clear that if an applicant takes a free phone, the fees in the second year are higher because of that free phone.  Shouldn't applicants have to cost-allocate by spreading the cost of the phone over the term of the contract?  So it would seem that even if you got the free phone under the bundling loophole, it's going to cut your funding in FY 2015-2016.

I can't get too worked up about it, though.  It seems clear that in an effort to free up money to provide more bandwidth and Wi-Fi funding to applicants without increasing the size of the fund, the FCC is following my suggestion to kick cell phones out of the program (and get rid of voice service in general).  So it's a little late to be making a decision about free cell phones when you're kicking cell phone service out of the program.  Cell phone service is an empty barn, so who cares if the barn door is closed.

Wednesday, May 21, 2014

Oh, it's warranted, all right

Remember the heady days after President Obama's ConnectED speech? Senator Rockefeller was talking about putting an extra $billion or so into the fund annually, and the administration was talking about putting an extra $5/phone into the program.

What happened to the idea that ConnectED (or E-Rate 2.0 or whatever) was to be an expansion of the program?  In his latest statement before Congress, Chairman Wheeler only touched on the level of funding by saying that the program must be "fiscally responsible."

The Chairman did recently say that he "will not hesitate" to increase the size of the fund "should this be warranted."  But then he goes on to explain why it's not warranted.

100 Mbps to every school and Wi-Fi in every classroom is going to cost something like $3.1 billion or $6.8 billion per year.  So if the FCC stands up to the howls from applicants and tosses everything else out and saves $600 million, we're still billions short.  Time to double the fund already.

Monday, May 19, 2014

Demand Analysis Analysis

Funds for Learning has released another E-Rate Demand Analysis, which is certainly worth a read.  The biggest change since last year: a new villain.  Last year it seemed like FFL was picking on the NYCBOE, making them their own category and showing how bloated their funding request was.  This year, they seem to have rural Alaska in the bullseye.  (Actually, FFL data analyses are usually balanced, so I assume that in reality, they were sifting the data and a few applications jumped out as extreme outliers.  As evidence that they don't have an axe to grind, they kept NYC as a separate category, showing that the Big Apple's request this year was quite modest.)

First, the silly complaints:
  1. "E-rate."  Even in the title of the document, which at least avoids the inconsistency others suffer.  If you're new to this blog, I hate the small "r."
  2. No pictures.  I have come to expect FFL reports to be filled with pictures so I don't have to actually read anything.  This report is just numbers.  Time to put on the waders.
Next, my analysis of the numbers.  And by "analysis," I mean "ill-considered comments on whatever jumps out at me."

Page 2: It's easy to see why FFL separated Alaska's data: the per-student pre-discount cost is $553, more than 5 times the next-highest state, and 10 times the national average.

Page 9: Look at the total per-student average of pre-discount costs: $36 for districts with a discount less than 60%, $47 for the 60-79% folks, $74 for those 80% and over.  I wish FFL had put 90% applicants in their own band, because, I think we would have seen an even bigger number for them.  So I'm back to my earlier rant on "incentives gone too far."  Why is there a discount matrix?  So that applicants with more low-income students have a more powerful incentive to purchase more.  Why?  I would say it's to ensure that low-income applicants can keep up with wealthier applicants.  But what these data show is that the incentive works too well: high-discount applicants spend twice as much as low-discount applicants.  Let's cut the top discount rate below 80%.  How about 65% like over at Rural Health Care?

Page 20: Suburbs request about 25% less per student than other applicants.  This reminds me of ESH's argument that applicants should be encouraged to get fiber, because it is cheaper and faster.  It makes just as much sense to say that applicants should be encouraged to become suburbs, because bandwidth is cheaper and faster in the suburbs.  That incentive would be easy to do with a new discount matrix: if 10% of your kids are low-income, you get a 40% discount if you're Urban, 50% if you're Rural, and 60% if you're Suburban.  Let's reward cost-effectiveness.  And hope that David Byrne was alone in thinking, "I wouldn't live there if you paid me."

Page 23: Holy crap!  Those remote Alaskans are spending almost $1,900/kid!  That's about 34 times the national average.  Remember when Obama said he wanted 99% of students to have high-speed broadband?  Well, I think we found that other 1%.

In short, my takeaways are:
  1. 80% off is too close to free.
  2. Throw Alaska under the bus.
  3. Fund the suburbs!

Friday, May 09, 2014

So long, voice, don't let the VoIP hit you on the way out

Driving my daughter to school this morning, I thought of another problem with trying to keep VoIP eligible if traditional voice is removed from the program.  It's cell phones.

It's pretty clear Chairman Wheeler wants to toss traditional voice out of the program: schools need to "evolve from analog voice to digital voice services."  It's not a question of whether to get rid of voice, but "What’s the best way to phase down support for legacy services like voice...?"  It's not so clear whether VoIP will get the heave-ho.

I'm in favor of getting rid of voice, too, especially cell phones, but I don't see how you can craft a loophole for VoIP.  Chairman Wheeler tries to draw a distinction between "traditional" and "digital" voice.  That's a might blurry line in general, and is even blurrier for cell phones.  Cell carriers are migrating to LTE for voice calls, and LTE is an IP-based protocol.  So some cell calls are VoIP or more accurately VoLTE-wioIP (Voice over LTE--which is over IP), and eventually all calls will be.  If you bought your phone recently, you have no control over whether your voice calls ride over LTE (VoIP) or 3G (circuit-switched).

Let's get VoWhatever out of the program.

Thursday, May 08, 2014

Wheeler's first wave

The first wave of Funding Year 2014-2015 is on the way.  Chairman Wheeler gave us a little preview in his speech at the E-Rate Modernization Workshop on Tuesday, and today we get the official announcement.  Time to update the table I posted back when the PIA procedures were approved.

Window close Days passed PIA approved Days passed First FCDL Total days Millions Applications
3/26/2014 -19 3/7/2014 69 5/15/2014 50 $607 14,600
3/14/2013 57 5/10/2013 19 5/29/2013 76 $130 12,023
3/20/2012 35 4/24/2012 77 7/10/2012 112 $646 23,800
3/24/2011 75 6/7/2011 19 6/26/2011 94 $398 18,500
2/11/2010 91 5/13/2010 13 5/26/2010 104 $429 18,200
2/12/2009 49 4/2/2009 26 4/28/2009 75 $134 6,931
2/7/2008 63 4/10/2008 21 5/1/2008 84 $352 10,000
2/8/2007 4/23/2007 74 $202 21,000
2/16/2006 4/26/2006 69 $184 4,880
2/17/2005 6/27/2005 130 $342 7,700
2/4/2004 4/27/2004 83 $43
1/16/2003 5/1/2003 105 $230
1/17/2002 4/24/2002 97 $233 9,300
1/18/2001 7/23/2001 186 $478
1/19/2000 4/14/2000 86 $253 13,000
3/11/1999 7/13/1999 124 $116 6,000

11/23/1998 $73 3,000
[See notes on table below.]

Chairman Wheeler did his best to make the first wave seem unprecedented, but it's in the same ballpark as other recent first waves.  But I have to say, it is an all-star first wave.  Why?
  1. It's the shortest time from close of the window to first wave.  By a long shot.  That's very good.  Far from my dream of a wave inside the filing window, but a good start.
  2. It's the second-largest first wave, behind only the shameful 2012 first wave, which came after the start of the funding year.
It's also exceptional in that it reverses the recent trend of processing simple applications first.  I guess it's part of the current Commission's irrational love affair with consortium applications.  I have to say, though, that the pendulum had swung to the extreme of getting simple apps out the door before looking at more complex apps.

So I'll put this wave in the win column for Chairman Wheeler.  It's not awesome, but it is progress.

Notes on table:
  1. The amount approved in the first wave is generally taken from the Search Commitments tool, which gives a higher amount than the announcements of the first wave.
  2. Anybody have the dates PIA procedures were approved before 2008?  I couldn't find them easily.

Tuesday, May 06, 2014

Nothing plus nothing

At today's E-Rate Modernization Workshop, I heard another justification for consortia: "aggregation of expertise."  At first blush, it seemed like a good reason to create a consortium, but then I thought about it.

Technology expertise does not aggregate all that well.  I thought of all the libraries in NJ: there are hundreds of them.  In general, they do not have a technology professional on the staff; when your total staff is 10, you have 15 computers and one access point, it doesn't make sense to pay for a full-time tech person.  And since few of them have many locations, there is even less expertise on wide-area networking.  If you excluded the 10 largest library systems, the aggregate wide-area networking expertise of the remaining libraries would be zero.  NJ has a statewide library network (kind of) because of the expertise of a single person at the State Library.  (With a little help from an E-Rate consultant.)

We did hear a couple of stories of statewide networks which probably got lower pricing because of their expertise in both technology and the marketplace.  (We also heard from one consortium that had to break up in order to get the best pricing, but that belongs in a different rant.)  But the expertise was not aggregated.  Those networks leveraged the existing expertise of the state agency.  I'm not just splitting hairs here.  There are two reasons the difference between aggregating and leveraging is important:
  1. "Aggregating" makes it sound like everyone pitches in.  In fact, the state agency pitches in, and everyone benefits.  Some organization has to supply the services of a very expensive asset (the expertise required to engineer a WAN is not cheap).  A state agency may be willing to do that for free, and some consortium leads have figured out ways to get a percentage to cover the cost of that asset, but if a bunch of applicants want to band together, one of them has to share a costly asset for nothing.
  2. And that assumes that one of the members had such a person to share.  If you aggregated the expertise of all the charter schools in NJ, you would not have the expertise necessary to get good pricing on a network to connect them all.  If you bring together 100 people who know nothing about WAN engineering and costs, the aggregate is zero.  You can't create expertise by aggregating the opinions of non-experts.  The kind of technology know-how you need to cost-effectively bring bandwidth to a consortium is not something you can get from crowdsourcing.
I'll use an analogy that toots my own horn, if you don't mind.  I am a Certified E-Rate Management Professional, and I've submitted (or overseen the submission of) hundreds of E-Rate applications.  Because this is my full-time job, I spend time on things like today's workshop, and reading every FCC appeal decision.  Here's my boast: if you aggregated the E-Rate expertise of all the employees in all the school districts in NJ, it would not equal my expertise.  600 people who each spend maybe 20 hours a year on the E-Rate just don't have the expertise of one person who has spent 2,000 hours a year on the E-Rate for over 10 years.  By the way, I wouldn't make that boast about PA or NY, because the Philadelphia and NYC school districts probably have people who spend 2,000 hours per year on the E-Rate, so it would at least be a contest.  (There was a woman from Philly at the workshop today, and she clearly understands the E-Rate really well.)

So I'm sticking with my position that the FCC should not encourage consortium purchasing.