In an earlier blog, the ordering process of organizations was explained in which different variants of making and recording orders and receipts are mentioned. Subsequently, it was explained how purchase order matching can speed up the process of checking invoices against orders and receipts. In this blog we will go even further into this topic and provide an explanation of the different ways in which invoices can be matched with purchase orders, namely automated on the basis of one order = one invoice, automated on the basis of line matching or a manual check of the order(s) and invoice. All three variants are explained in more detail below.
One order = One invoice | Match on total amount
The principle of one order = one invoice is based on the fact that one invoice relates to one purchase order and that the total amount is matched. This form of matching assumes that the total amount on the invoice must correspond to the total of the booked receipts on the purchase order, without looking at the details on the individual lines of both the order and the invoice. In practice, this often means that the supplier will send more invoices with less content. There is an invoice per order or sometimes even per delivery.
The method one order = one invoice assumes that a match always takes place at the moment:
• The prices of the items in the order are the same as what is on the invoice;
• The supplier invoices for the sames amount of quantity of the order as the recipient has recorded (as received) in his warehouse,
On the other hand, there will be no match between invoice and order at the time:
• The price of an item changes between the time the order is created and the final receipt of the invoice;
• Not all items on the order are accounted for (but appear on the invoice);
• There is a discrepancy between the items invoiced and the items delivered from the suppliers.
If one of the above scenarios applies, the invoice is usually referred back to the responsible buyer, who checks the order and invoice and contacts the supplier to resolve the difference together. This often comes down to a subsequent delivery, an adjustment of the price in the order or a credit note from the supplier.
With the method one order = one invoice, it is assumed that there should be zero probability for any discrepancy between the amount on the invoice and that on the purchase order (and receipts) and that they are equal. While at line level there may be some deviations that would cancel each other out in the end, not affecting the total amount which remains equal. For instance, if ten different items are ordered in the same numbers (quantity) and at the same price (eg €1), the chance is very small that the price of one item will be € 0.90 and that of another item € 1.10, however, should this occur the total amount remains the same as difference would cancel each other out.
For matching invoices on the basis of one order = one invoice, “standard OCR invoice recognition” is sufficient, whereby only the header data of the invoice are recognized, such as supplier, invoice number, invoice date, order number and total amounts. The minimum recognition that suppliers realize on this header data is often 80%, whereby this recognition can often be trained to more than 95% recognition. The high degree of recognition is a major advantage of this variant of matching.
Rule matching | Match at line level
A more extensive variant of purchase order matching is order detail or line matching. Matching an invoice at line level actually means that an invoice based on numbers and unit prices (at line level) is matched with the previously created purchase order and the receipts booked from the ERP system. With this form of matching, an invoice is automatically compared at the detail level (article number, quantity, unit price, etc.) with the original purchase order, whereby every difference is recognized and appears as ‘error’ if one of the characteristics deviates between invoice and order. This detailed way of matching basically ensures that the margin of error that the matching on total amount entails is kept to a minimum. After all, every detail on the order is compared with the same details on the invoice.
For matching invoices at line level, it is very important that the line details are also recognized by the OCR software. In addition to the header data, at least the order number, item numbers, numbers and unit prices must also be recognized at line level. Because more data from an invoice has to be recognized, the basic recognition is in many cases around 50%, which can then be trained to increase to approximately 80%. It should be noted that you do not know in advance which 80% is recognized correctly and in many cases users choose to go through each field manually to be sure that the recognition went well. The recognition software that is required for line matching also has a considerably higher initial investment than recognition software that only recognizes the header data of an invoice.
An advantage of line matching is that multiple order numbers of one invoice can be recognized, which can then be matched in one go. If a match is found between one invoice with multiple purchase orders, the different order lines are retrieved in the ERP system and the invoice can be posted automatically in the financial administration.
Another advantage of line matching is the possibility to have a different line come up immediately. In invoice approval, in some cases it is also possible with line recognition to present the errors at line level as such for verification to the responsible buyer / budget holder in the organization, so that this person no longer has to search for which item line a discrepancy was found.
With line matching, however, the challenge lies in the exceptions that often occur when multiple purchase orders have to be matched on the same invoice. Think of scenarios in which multiple orders must be matched with an invoice and not all invoice lines correspond to one of the previously made purchase orders. Automating these exceptions makes line matching even more expensive than usual because more extensive and therefore more expensive OCR recognition is required.
For some organizations, matching invoices with purchase orders based on detailed rules or total amounts is (technically) not possible or not desirable. An example of this is organizations that keep track of their orders in an Access database, Excel list(s) or a custom-made system with which an external (software) supplier cannot link. In general, the following rule applies here: If a purchase order is not recorded in the same system where the invoice is recorded, then it is difficult for both to be automatically linked to each other, let alone automatically matched and posted.
If it is technically feasible to link systems, a financial consideration can of course also play a role in whether or not automatic matching is being developed. The alternative is often to opt for a manual (optical) match, in which the human eye actually determines whether or not the invoice matches the order.
Although invoices in such a situation are not automatically posted if they match the purchase order(s), there is still a lot of benefit to be gained for these companies should they digitize their invoice flow. Many organizations, for example, choose to make both the order and the packing slip digital and attach it to the invoice as an attachment, so that one digital archive is created where all documents can be found together.
If we look at the different forms of purchase order matching, we can draw a number of conclusions. Above all, one method is not necessarily better than the other and it is up to each organization to choose one of the three forms described above.
The first preference of a potential customer often goes to rule recognition, because this is emotionally closer to the current method. Many invoices currently entering the organization contain multiple lines, whether or not broken down into different purchase order numbers. A collective invoice was always very useful, because the rule of thumb in many companies is still “the fewer invoices we get, the smaller the chance that an invoice will be lost”. From that perspective, line matching often fits well with the initial needs of a company that is looking for a solution for the digital processing of purchase invoices. In fact, one does not have to change much in one’s own working method to be able to get started with digital invoice processing in general and purchase order matching in particular.
The question is, however, whether a company is really helped in the long term with line matching, because the exceptions occur more often than previously expected and moreover take more time to handle than previously thought. In more detail: the more orders / deliveries there are on an invoice, the greater the chance of a deviation, the probability increases that more people may need to look at an invoice to resolve this deviation. In this way, the chance is greater that an invoice, despite digitization, has a longer handling time.
What software for matching invoices with purchase orders cannot do is resolve differences in data (between order and invoice for example). The software only shows the difference at all times. If we take this into account, then it is much more obvious to make purchase order matching as small as possible and therefore to match as few orders as possible with an invoice. The fewer orders there are on an invoice, the greater the chance that an invoice matches and the higher the matching rate, therefor lowering the human action needed to review these invoices.
Moreover, the principle of one order = one invoice is less complex to automate, although it often requires a (small) change in the current working method. Suppliers must invoice per order or sometimes even per delivery. This is generally not overnight, and therefore sometimes you have to accept (temporarily or otherwise) that some suppliers continue to send a collective invoice with multiple orders that is not automatically matched. However, many suppliers are quick to comply if you tell them that a collective invoice with many orders, in which one order deviates, is in many cases not paid until the deviation is resolved. An invoice with one order will match more often and will therefore be paid on time regularly. For the speed of processing, it is also very practical to immediately ask the supplier whether he can send his invoice digitally or electronically from now on. This makes processing the invoice much easier for both of parties; no more folding invoices into envelops or stamps needed for the supplier and as a customer you no longer have to scan the paper invoices.
Finally, we see in practice that organizations have a hard time achieving a high matching percentage with one order = one invoice, let alone matching multiple orders with the same invoice. The principle of one order = one invoice ensures that the process is considerably simplified and becomes very well-organized and clear, both for you and for the supplier. That is precisely the starting point that we would like to present to our customers the moment they start digitizing and automating their administration.
In theory, every invoice should of course correspond to the orders that have been made. Whether there are one or 100 orders on an invoice, an agreement is an agreement and if each party fulfil their agreements, it should always go well. That is the theory. The practice is a bit different and that is why we have written a blog in which we provide you with tools and tips to benefit the most from incorporating purchase order matching.