VII. Capability of Verizon's Hot Cut Processes to Meet Demand 47 C.F.R. § 51.319(d)(2)(ii)(A)(3)
A. Position of Verizon
Verizon argues that its BHC process is scalable to handle the greater volumes of hot cuts and related work that would be expected from the elimination of switching as a UNE offering. Verizon proposes to meet the increased demand by increasing the size of the work force at its central offices and other work centers. Verizon developed a spreadsheet model referred to as the "Force-Load Model," which first determines the incremental level of hot cuts and winbacks that would be required in a post-UNE-P world. The work volume assumptions and data are sponsored in testimony of Dr. Taylor. The Force Load Model then converts the incremental work requirements into staffing level increases in the central offices. Similar techniques are used to determine increased staffing level needs at other Verizon work centers.
For purposes of analyzing the volumes entailed in migrating the embedded UNE-P base to UNE-L, the Force Load Model segments demand into five time periods. The UNE-P embedded base conversion is assumed to be completed at the end of 27 months. Thereafter, only the only incremental demand would be caused by customer migrations. Verizon contends that it can hire and train any additional work force needed in a relatively short period of time
B. Positions of Other Parties
AT&T, MCI, CalTel and Covad take issue with Verizon's claimed processes, arguing that Verizon's processes exhibit similar flaws to that of SBC. Although the current volumes of UNE-P and UNE-L in Verizon's territory are much smaller than those in the SBC territory, the parties claim that Verizon's proposal is not sufficient to meet the expected CLEC demand if UNE-P is eliminated. The parties criticize Verizon's assumptions for not being forward looking. MCI argues that Verizon's past performance with small volumes is no indicator of how its process might perform under dramatically increased volumes. AT&T further argues that Verizon's actual base of UNE-P loops that will have to be cut is roughly double what Dr. Taylor projected using historical data. AT&T bases this claim on the position taken by Verizon's appeal of the Commission's UNE rate order, filed in the U.S. Court of Appeals, in which Verizon claims that it has lost 80,000 local customers to UNE-P migration and such orders have escalated to 35,970 per month, a 2200 % increase since January 2003.53
Verizon also makes inconsistent assumptions concerning expected hot cut volumes in its TELRIC pricing analysis versus in its scalability analysis. Verizon assumed greater amounts of work time in its TELRIC analysis as compared to its Force Load Model which was used to support claims of the adequacy of its existing work force.54 Although Verizon claims that the two sets of estimates are consistent "at the aggregate level," Verizon fails to show how specific inconsistencies in the two sets of estimates, such as those identified in AT&T witness Falcone's testimony, can be reconciled.55
Another issue relating to the scalability of Verizon's processes has to do with the manual nature of the work. MCI claims that Verizon's process requires various manual tasks and that Verizon cannot handle the volume of hot cuts necessary to ensure a seamless, timely and reliable migration of customers among carriers in the absence of UNE switching.56
Verizon's migration analysis neglects certain types of UNE-P lines, such as digital and hi-cap lines, even though carriers using UNE-P to provide these services would be required to convert them to alternative facilities in the same 27-month timeframe.57 Under Verizon's scenario, all the embedded UNE-P lines are to be converted to alternative provisioning methods. This produces an understatement of the embedded UNE-P base and leads to an understatement of the monthly conversions of this base during the transitional period.58
MCI also challenges Verizon's "steady state" concept, which reflects the period in which the market matures and demand becomes relatively constant. Verizon estimates that UNE-P migrations would reach a stable level by the 8th month of the 27-month conversion period. During this same period, Verizon calculates winbacks by applying a percentage to the number of UNE lines, resulting in a substantial increase of monthly winback volumes. Thus, MCI contends that Verizon's study predicts decreasing CLEC demand in contradiction to the steady state assumption. In order to properly model a steady state, UNE-P migration counts need to grow to compensate for the loss of CLEC customers to Verizon, disconnect and other causes that were excluded from Verizon's study.59 Therefore, to meet these net growth volumes, MCI argues that UNE-P migrations should grow each month (starting in month 9) by the amount equal to the total losses in CLEC lines, which are proportional to the total stock associated with former UNE-P lines.60 As a result, MCI contends that proper modeling of Verizon's steady state process increases the volume of hot cuts associated with migrations by a factor of 1.6 in the 27th month of the transition period.61
In Table 3 of its Opening Brief, MCI provided a comparison of Verizon's volume study assumptions versus MCI's assumptions. MCI's adjustments affect the levels of net growth and total migrations assumed starting from month nine. Winback volumes increase due to an increase in the stock of UNE-L lines.62 Migration volumes increase to account for the loss in CLECs' lines due to winbacks and disconnects.63
The impact of MCI's volume adjustments on the results of the Force Load model is summarized in Table 4 of MCI's opening brief. MCI's assumed hot cut volumes would increase the number of additional employees required to support migration-related hot cuts. Work force support increased starting from month 9: by a factor of 1.03 and by a factor of 1.6 in month 27 compared to Verizon's study. Because MCI made no adjustments to the embedded volumes, the support of the embedded conversions is the same level as in Verizon's study.
MCI also argues that Verizon ignored two components that its own volume witness, Dr. Taylor identified as incremental hot cuts: (1) migrations from UNE-P to Resale and (2) from UNE-P to UNE-L.64 Dr. Taylor admitted that the first component should technically be counted towards total incremental hot cuts, but argued that the volumes for this component are insignificant because CLECs' share in California is relatively small.65 MCI contends, however, that the migration volumes ignored by Verizon's volumes study - due to former UNE-P lines migrating to resale and other CLECs - would also increase dramatically compared to the current volumes.66
MCI also argues that Verizon excludes disconnects from its migration study.
At the time a CLEC places an order, Verizon won't know when the central office will reach "critical mass," and thus when the hot cut will occur. Verizon offers to give CLECs six days notice, unless the CLEC has not received a hot cut date by the 20th day after placing an order; then the CLEC will know the cut is planned for the 26th day. MCI argues that such uncertainty creates administrative difficulties, inefficiencies and customer satisfaction problems. MCI also objects to Verizon's proposed holding time for BHC orders as being too long to meet CLEC customer needs and proposes that orders be queued no longer than seven days. Embedded UNE-P customers being migrated to UNE-L must be notified so they can re-program dialing features supported by the carrier's switch. Thus, customers moving from one carrier to another are affected by lengthy provisioning delays.
Verizon agrees to consider establishing provisioning intervals for each central office, but only after the Commission approves its proposal. MCI argues, however, that specific intervals for each of Verizon's central offices must be established before approval of a batch hot cut process, or there is no guarantee it will ever happen.
Verizon also offers to provide CLECs a "UNE-P like" service during the lengthy wait for a customer's loop to be migrated to CLECs' switches. MCI argues that Verizon's offer is unacceptable because it will not be priced the same as UNE-P (i.e., likely higher), and it will lack the protections applicable to UNEs, thus likely increasing a CLECs' costs of migrating customers and subjecting them to discriminatory or other unacceptable practices. MCI thus opposes Verizon's proposed "UNE-P like" service and instead proposes that Verizon implement a batch hot cut process capable of handling the necessary volumes without imposing excessive delays on CLECs.
Without UNE-P, CLECs will have to put in place necessary equipment, transport and collocation arrangements in order to migrate customer loops to their own switches. Building or expanding a collocation cage, building backhaul facilities, expanding Verizon tandem transport trunks, and expanding tandem switches takes several months.67 These time-consuming preparations make it likely that CLECs' migration will be in peaks and valleys, not a smooth line.
Verizon also claims that the actual number of hot cuts will be lower than its estimates because some customers will choose alternate technologies and bundled voice and data services will decrease customer churn. However, MCI disputes there is any evidence that customers will choose to leave the telephone network. Further, Verizon refuses to support migrations of CLEC customer loops with bundled voice and data - the very type of customer that Verizon claims will reduce the amount of churn - in its batch hot cut process.
CalTel witness Compton testified that based on Telscape's actual experience to date, Verizon's systems for transitioning customers from UNE-P to UNE-L are not fully scalable or robust enough to migrate the quantities of existing UNE-P customers to UNE-L that would result from elimination of UNE-P.68
AT&T proposes that Verizon be required to propose a volume scalability test of its BHC process within 30 days of an initial Commission decision and to conduct that test and provide results to CLECs and the Commission. AT&T proposes that the Commission not give final approval to Verizon's BHC process until Verizon has successfully completed the scalability test.
C. Discussion
As discussed above with respect to SBC, we conclude that doubts remain as to whether Verizon's proposed volume limits based on accumulating a "critical mass" of orders will be sufficient to meet increased demand for hot cuts that would result with the elimination of UNE-P. Actual volume limit requirements will depend upon the nature, extent, pace, and allotted schedule for the transition from UNE-P to UNE-L. We agree with the concerns raised by MCI and AT&T with respect to the risks of errors and delays that would be faced by Verizon in meeting its volume commitments.
We agree with the adjustments that MCI calculated concerning expected growth in hot cut volume that would be required to be consistent with the "steady state" assumed in Verizon's Force Load Model, assuming a 27-month transition period as originally prescribed in the TRO. MCI's adjustments to Verizon's hot cut volumes result in an increase of the number of additional employees required to support migration-related hot cuts: work force support increased starting from month 9: by a factor of 1.03 in month 9, and by a factor of 1.6 in month 27 compared to Verizon's study. Since MCI made no adjustments to the Verizon's embedded volumes, the embedded conversions assumed in Verizon's study are not changed.69
Also, Verizon has not taken into account the potential growth in hot cut volume that could result over time from incorporating additional migration scenarios into the cutover process, as discussed below. Therefore, we agree performance measures and testing in some form are appropriate to provide reasonable assurances that Verizon will be able to meet necessary volume demand for hot cuts with the elimination of
UNE-P.
D. Verizon's Provisioning Intervals (47 C.F.R. § 51.319(d)(2)(ii)(A)(2))
Verizon's manual processing of orders for the hot cut process also creates difficulties with respect to provisioning intervals. Verizon requires a 15-business day provisioning interval for a large job and between 6 and 26 business days for its proposed batch process, depending on when a "critical mass" of orders is reached in a given central office. AT&T witness Falcone testified that as a practical matter, a CLEC cannot use the Verizon large job and batch hot cut processes as customer acquisition tools, because of the delay between the CLEC order date and the hot cut execution date. The large job process was specifically designed to cut over a large quantity of lines only after customers have been acquired by the CLEC. The batch cut process is likewise not practical because the CLEC must be able to give its customers a date certain when the CLEC can migrate their service over to the CLEC switch. Given Verizon's provisioning interval between 6 and 26 business days, however, the CLEC cannot provide a date certain. Instead, CLECs must use Verizon's "UNE-P-like" service to acquire customers, and then issue a second order to move the customer to the CLEC switch using the batch process. We thus consider Verizon's provisioning intervals to be deficient. We shall require Verizon to revise its processes to provide the capability to offer specific provisioning intervals so that CLECs can inform their own customers as to a date certain when the cutover will occur.
53 See Ex. 157 , Verizon's Appellate Brief at 9
54 Ex. 155C, Falcone Testimony, Ex. RVF-4, at 36:1-38:5
55 Ex.155C, Falcone Testimony, Ex. RVF-4, at 39
56 Ex. 143 Lichtenberg/Starkey, at 67-68.
57 Ex. 143 (Lichtenberg/Starkey 1/15 reply), at 71-72.
58 Ex. 143 (Lichtenberg/Starkey 1/15 reply), at 72.
59 Ex. 143 (Lichtenberg/Starkey 1/15 reply), at 75.
60 Ex. 143 (Lichtenberg/Starkey 1/15 reply), at 76.
61 Ex. 143 (Lichtenberg/Starkey 1/15 reply), at 74
62 Ex. 143 (Lichtenberg/Starkey 1/15 reply), at 78.
63 Ex. 143 (Lichtenberg/Starkey 1/15 reply), at 78.
64 Ex. 102 (Taylor 11/7 Direct), at 16, Table 3.
65 Ex. 102 (Taylor 11/7 Direct), 26, n.7.
66 Ex. 143 (Lichtenberg/Starkey 1/15 reply), at 70.
67 Tr. Vol. 54; 1/29/01; 8301:19-22; 8366:6-8 (SBC witness cross-examination agreeing that it could take 120-180 days to provision cage-to-cage cabling for CLEC line splitting and months to provision a collocation cage).
68 Ex. 180 at 17 (Direct Testimony of Compton). Compton testified that Telscape is the largest user of UNE-L in California, having about half of the UNE-Ls installed in California.
69 See MCI Opening Brief, pg. 267
