January 3, 2008
The Matchback Process


Wendy received a children’s clothing catalog and wanted to order a snowsuit for her daughter. Wendy always orders online, so she went to the website and started to look up the snowsuit she had seen in the catalog. Unfortunately, she left the catalog on the kitchen table and did not know the item number or any of the codes printed on the catalog. She ultimately purchased a jacket and snow pants that were not shown in the catalog, and a hat and glove set that was in the catalog. She wanted the items for an upcoming trip to her in-laws’ home in snow country, and had the items shipped to herself at her in-laws’ address. Needless to say, she did not enter a key code at checkout.

This story is not atypical for catalogers these days. How do multichannel marketers determine what contact vehicle gets credit for a sale? How do they determine who the customer is?

The matchback process is the science and art of allocating sales to the appropriate customer and contact vehicle. Once more science than art, when customers called or mailed in their orders, matching back sales accurately was simple. Now it has become more challenging: to blend technology and marketing with the goal of statistical validity.

Because marketers want to be able to create more effective and efficient marketing plans with limited marketing budgets, and ensure a high degree of confidence before implementing costly testing and roll-out strategies, they must be sure to:

  • Allocate demand to appropriate campaign/promotional effort/marketing strategy (e.g. catalog/phone, paid and organic search, affiliates, email, bounce backs, retail, other non-mail).
  • Allocate accurate marketing costs to each order since variable costs continue to rise.
  • Show ROI accurately by campaign/promotional effort/marketing strategy.

The matchback process seeks to understand a customer’s path to purchase by source and channel. In order to be most effective the process must:

  • Determine the most accurate allocation to specific contact vehicle(s) that triggered sale.
  • Create a method to allocate credit for a sale among multiple contact vehicles as appropriate, to understand the number and mix of contacts.
  • Tailor the process to the specific title’s business needs and practices; each title has unique characteristics (contact strategy, merchandising, and business objectives).

Today, industry estimates indicate as many as forty percent of mailers do not do matchbacks at all. And those who do employ the practice have a variety of methodologies from matching back by each channel separately to complex fractional analysis. The generally accepted rule is to match back to the most recent contact or most recent catalog depending on business objectives. As seen in the vignette above, it may not always be so easy or clear cut to determine even what the most recent contact was.

Best Practices

We offer the following advice:

  1. Be obsessed about capturing source codes on the front-end to minimize the unallocated pool. Make sure all customer service representatives understand the need and are trained to ask for the code whenever an order is placed. Make it easy to enter the code at check-out on the internet.
  2. Include data from at least one year when starting up the matchback process; for updates, include new orders since the last update.
  3. Determine how frequently to perform matchbacks based on how they are used; factor in the update schedule and a cost-benefit analysis. If you are mailing quarterly, you may not need monthly matchbacks.
  4. Include all orders from all channels in the business rules; offers and order channels are interrelated. For example, do not automatically eliminate an order from the process because it came from an affiliate or paid search. Affiliate sites offer benefits that may encourage customers to place orders through the site even though the purchase was driven by a catalog. Paid search terms, especially product or brand-specific ones, could easily be catalog or email-driven (but the catalog is on the kitchen table).
  5. Use all the data available to build business rules for the process:

    1. Promotion/contact history (mail/contact file).
    2. Other segments (requestors, gift recipients, web only responders).
    3. Circulation plan.
    4. Responder file (transaction data).
    5. Non-mailed source codes (package inserts, paid search, affiliates, bounce backs).
    6. Campaign metrics (order curves).
  6. Validate logic quarterly using percentage thresholds to identify arenas and rules for deeper investigation.
  7. Perform hygiene on mail file and transaction data before matchback process begins to have the most accurate file and to increase matches.
  8. Consider item level codes to be able to identify results at SKU level when source codes are not available, in order to determine best contact when more than one contact vehicle is live.
  9. Make sure all active contact vehicles are included in the process; use order curves to determine appropriate time frames for inclusion.

The methodology will not be the same for all marketers. Business goals and marketing objectives will shape the process. However, what will most likely be true for all marketers is that some sort of fractional allocation will be necessary. Customers are exposed to multiple contacts and are potentially influenced by each of them. Each company must determine channel-specific weighted business rules based on the pool of potential matches, all open campaigns, the value of the marketing program to the business, and the reach of the campaign. Historical performance analysis can be used to create hypotheses for channel weighting. Testing is difficult since the control group cannot be isolated from all possible contact vehicles or venues.

Traps to Avoid

The following practices will distort results. Be careful NOT to:

  • Overstate email or online performance which can lead to lowering circulation misguidedly, such as:
    • Eliminating profitable segments
    • Removing drops completely
  • Match back orders from different channels within their own siloes; make sure all orders are included in the business rules for matching back.
  • Match back to support a specific business objective rather than matching back to present as accurate as possible picture of customer’s path to purchase (e.g. matching back all orders possible to most recent catalog, ignoring email contacts).
  • Fail to review business rules regularly (we recommend quarterly); matchback rules need continuous improvement.

By accurate and detailed application of coding circulation and merchandise, being attuned to the complexities of contact vehicles, and a constant review and honing of rules and practices, multichannel marketers will be able to gain a more complete picture of the success of individual campaigns and therefore be able to allocate resources more effectively.

We're interested in hearing from you.
Contact us