Meeting the Challenge of Wholesale Carrier Reconciliation

Automation of wholesale invoice auditing is a challenge facing telecom operators. Audits are necessary because 80 percent of the invoices are incorrect. Take for example the case of international calling, where the market has become a commodity and competition among the international resellers is driven by price. Any reseller who can cut an underlying cost by up to 10 percent will have a competitive advantage because they can offer lower prices to their customers. Effectively auditing the wholesaler's bill can help this competitive advantage become a reality.

In the past, most carriers didn't even maintain a job description that included finding errors in the charges on carrier invoices. An accountant was supposed to analyze the correctness of the charges before submitting payment for approval. In most cases the bill was checked to see if it "looked right." In the best case scenario, the management summary page of an invoice listing number of calls, number of minutes and the corresponding charges was compared with each rating area. But to get a comprehensive view of the accuracy of the invoices, the following questions must be answered:

* What are the major reasons for overcharges?
* How do you determine if the carrier used the correct rates?
* How do you match your own call detail records with those of the wholesaler?
* How do you identify duplicate billing? Non-billable calls?
* How do you receive credit for poor service from a carrier?

Overcharges

The major reason for billing overcharges is that the wholesale carrier applies rates are that higher than the contractual rates. This can sometimes account for more than 60 percent of overcharges. Second, in an international reseller environment, the reseller often has a tax-exempt status. Often, however, the wholesaler will not acknowledge the tax exemptions. In such a situation, the reseller needs to manually review each invoice for applicable taxes and send a credit note for taxes that were charged incorrectly. Finally, there are usually discrepancies in circuit installation or disconnection dates. These dates need to be verified as to when the circuits were set up or disconnected. Discrepancies can be resolved by manual inspection of dates from the circuit records that the CR department maintains.

Incorrect rates

No carrier has a flawless call rating record all the time; most carriers make call rating mistakes. Carriers that use aggressive pricing strategies tend to change their rates very frequently, so bi-weekly changes are quite common. With such rapid changes, it is nearly impossible to keep rating databases accurate at all times. New rates are entered too late or too early, and the large number of changes increases the likelihood of errors. When wholesalers change a rate before a billing cutover date, they rarely make an adjustment in their billing program to apply both rates by prorating each part according to its contractual effective date and time.

Usually, a reseller's contract will specify start date, end date, rates, billing increments and other billing information such as minimum billing duration, number of free directory assistance calls, and so forth. Contract amendments, mainly rate changes, will go through the reseller's legal department and then be added to the present rates. The reseller's databases must be maintained and updated with the new contract amendments and rates along with the time period for which the rate is effective. Calls then can be re-billed using carrier CDRs to verify whether the rates used by the carrier were correct. Once the differences are identified, the wholesaler can be informed. To be able to do this, it is very important that the reseller maintain updated rate tables.

Incorrect recognition of rating areas

Explosive growth in the number of area codes (NPAs) and NXXs within the North American Numbering Plan, as well as changes in international numbering make the maintenance of the telephone numbering tables a true nightmare.

These tables are used to determine the rating areas for the calls. At times, a carrier may use only a three-digit code to identify a calling area instead of a seven-digit code, resulting in incorrect use of rates. To give a recent example, one carrier recognized all international calls starting with 396 as going to Vatican City. However, country code 39 is for Italy and city code 6 is for Rome, not Vatican City. This carrier had different rates for calls to Italy and Vatican City and the latter were higher. The code that uniquely identifies calls terminating in Vatican City is 39-6-698 followed by a five-digit number. This carrier was using only the first three digits to identify the calls. Based on the digits 396, it identified all calls going to Rome as Vatican City and thus charged these calls at the higher rate. ITU already has assigned a new three-digit country code-379--for Vatican City.

To avoid this kind of problem the reseller must maintain an up-to-date database of location codes. The database must identify the location based on the first seven digits of the called area and not just the first three, as in the example above. Although these types of errors may not contribute to major savings, they could definitely cause major billing discrepancies if there is considerable call traffic to an erroneous code.

Incorrect rating due to discrepancies in time period recognition

If a call starts during one rate period and ends in another rate period, it should be billed at the rates applicable for each period. In a situation when a call spans multiple rate periods depending on the time of the day, it is necessary to break a single CDR into multiple CDRs (whenever a CDR crosses the boundary of a rate period). The multiple CDRs are separately computed and then the total cost is calculated. With competition, such multiple rate plans will become more common.

Matching your own call detail records with those of the wholesaler

At the first level of comparison, the total number of calls and the total number of minutes are compared, calculating the values from the electronic CDRs and comparing them to the paper invoices received from carriers. At the second level, the total number of calls and number of minutes for each rating area and rate time period are calculated. The calls and numbers of minutes are grouped by the rating areas and time periods applicable to the invoice.

The third level involves call-by-call matching to locate each call from the carrier's CDRs and match it to a corresponding call record in a reseller's traffic log. To begin record-by-record matching, the origination number, destination number, duration and start time must be physically matched. In addition, the clocks in the two switches need to be calibrated, since they may be in a different time period. Several wholesalers provide start times in increments of minutes (not seconds); thus, call records from a reseller's switch need to be rounded off to the nearest minute before call comparison can begin.

During the call-matching process, all records may not exactly match. Thus, calls need to be compared in multiple passes, allowing for increased deviations of duration and start time in each subsequent pass. For instance, in the second pass the deviation of duration would be increased and start time to 30 seconds to try for a match. By going through this iterative process, calls that have fallen out of an acceptable range of deviation can be identified. By using this automated process, the following issues can be identified:

* Unmatched calls: Calls on carrier CDRs that cannot be found in the reseller's traffic CDRs. This is a special case of call duration discrepancies between carrier and traffic records where the call duration = 0 seconds according to the reseller's traffic records.

* Minimum billing increment errors: Actual minimum billing increments used by the carrier billing program are different from those specified in the contract.

* Duration discrepancies: Call duration according to a carrier is different from the duration according to the reseller's traffic records.

Identifying duplicate billing and non-billable calls

The wholesaler's CDRs need to be reviewed for duplicate billing, both within the same billing period and across overlapping CDRs in different billing cycles. By using an automated process, calls are identified in the carrier's CDR files to find instances of double or multiple billing. Programs can be used to scan each CDR file for duplicate records. In addition, when the call date and time ranges overlap in two or more CDR files related to the same account, the overlapping records need to be checked for duplicates.

When a reseller's rating-area recognition database identifies non-billable calls made to invalid numbers, such as an insufficient number of digits dialed to complete a call, or nonexistent numbers, called numbers need to be checked against a location code database to verify that the location is correct. In addition, the number of digits dialed needs to be counted to ensure that a valid number was dialed.

By matching the carrier calls to traffic CDRs, a reseller can locate calls that aren't found in traffic records. A high proportion of these unmatched calls may be non-billable due to the following reasons:

* Unsuccessful attempts to complete a call (essentially non-billable calls): very short duration calls followed by another call to the same number within 60 seconds.

* Very short calls (essentially non-billable calls): the call duration is so short that no billable activity could occur. Credit may be requested for all calls lasting between 1 and 10 seconds.

* "Hung trunk" (also called "stuck clock") calls: actual call duration was shorter than recorded as a result of failed carrier disconnect supervision.

* Poor-quality calls: reported by credit manager and trouble tickets.

* Test calls: reported by trouble tickets.

* Prematurely disconnected calls: reported by credit manager.

* "Non-connects" (uncompleted calls): reported by credit manager.

* Fraudulent calls: these are often the cause of contention between reseller and carrier since they are hard to identify.

* Calls in traffic CDR matching specific criteria for special projects.

Receiving credit from wholesale carrier for poor service

The trouble ticket department provides the carrier reconciliation department with a record of all trouble ticket and test calls. Carrier reconciliation then sends credit requests for these calls to the carriers. The retail credit manager provides a list of all poor-quality calls reported by retail customers. Credit requests are processed for such calls and claims are then sent to carriers.

Conclusion

In general, anywhere from 8 percent to 84 percent of credit requests are accepted by carriers. In today's competitive world, this is surely not a figure that can be overlooked. In fact, this could be the difference between you and your competitor and provide you with the winning edge.

Ankur Lal is President of Infozech, a company specializing in billing and reconciliation solutions. He can be reached at ankur@infozech or via the company's Web site at http://www.infozech.com.