One of the greatest strengths of any form of database marketing, including the catalog arena, is that the numbers can be massaged in such a way that they will "talk to you." You want to find those numbers that help you move in the best direction and avoid critical pitfalls.
First of all, step back and ask a few questions. It is imperative to ask the right questions. Clarify what you need to know in order to make decisions for the future. Then, you can seek the corresponding data, measure it, analyze it and use it as input to the decision-making process. Be sure that you give priority to numbers that will enable you to take swift action; to understand key changes; give hints as to what areas should be evaluated more thoroughly "send up "flares." You may miss important insights if you only look at the existing measurement data. At the same time, there is a need to narrow down the data derived from a potential overabundance of numbers.
An entire book could be written on the hows and whys of measuring success. So, let's take a broader approach. Let's examine some key questions that should apply to any cataloger. They will give you a foundation to build a more complete set of questions that reflect the needs of your unique catalog situation.
#1: What Specific Goals Should We Have For Our Organization?
As in any situation, catalog success is best understood in context. Absolute numbers are only valuable when placed in context with past history and future expectations.
To be meaningful, success must be measured against goals, which have been based on past performance, current industry status, and perceived opportunities. Or, in the case of a new venture, a "best guess" set of goals based on similar industry performances and the requirements of funding the business.
Any form of measurement must start with a comparison against some type of "yardstick." For a catalog business, that tool should be a Strategic Plan.
In its simplest form, a Strategic Plan is a written document that answers the following questions "Where have we been? Where are we now? Where are we going?
The financial aspect of a Strategic Plan should be developed from the "Bottom Up." In other words, establish a circulation plan for the next three years. Apply costs and revenues, based on reasonable assumptions, including the effects of inflation and other outside influences. Be realistic, but include goals that will stretch the organization, so that complacency can be prevented.
Work with the initial plan, carefully thinking through various strategies. However, do not use every opportunity (e.g., media type, new venture idea) you can think of in the first cut. Compare top-line sales and bottom-line profits with any corporate goals that may have been assigned. Assuming that the first cut is inadequate, go back to the bottom again to stretch the original goals more than you did originally. Continue the process until everyone is satisfied that all aspects have been considered, including staffing, office space, warehouse space, and more. Every department should be involved in this process to the greatest extent possible. That way, there will be more ownership of the plan and less finger-pointing down the road.
#2: How Should We Look At Sales?
Certainly, total sales are meaningful. However, a relative measurement for sales helps bring the business into focus:
(1000 * (Response Rate)) * Average Order Size = Sales Per Thousand Catalogs sent
Sales/Book is often used, as well. The difference is the position of the decimal point and the choice is based on personal preference or intra-company tradition.
These relative numbers allow easy comparison and are simple to carry in your head. Thus, you will be able to calculate trends quickly and compare with past history. At higher levels of management, concentrate on concise summary data.
Sales can then be broken down in meaningful ways:
- Customer versus prospect "For most businesses, customer Sales Per Thousand are higher than prospects (i.e., first-time buyers). Customers should be generating profits even if prospects are acquired at a loss.
- Seasonal sales.
- Prospect source.
- Type of customer.
Sales should also be evaluated based on timing. Seasonal variations should be measured in absolute and relative terms. Also, response curves should be developed so that results to-date can be projected to their likely conclusion.
#3: How Should We Measure Profits?
Here is a sample profit and loss statement for a catalog. One of these should be developed for each event (e.g., Spring Prospect Catalog) and each test within an event:
Such budget line items as personnel salaries, benefits, agency retainers and other overhead should be handled at a corporate level. Many organizations do not include those here for the sake of simplicity.
There are certain decisions that should be consistent throughout the organization. For instance, determining the average cost of shipping out products. Those costs should be used consistently unless a change needs to be made for a particular test or event to reflect actual differences. For instance, in a shipping and handling test, a higher cost per order should be used for the test cells.
In evaluating tests, a preliminary P&L may be constructed with certain test assumptions to gauge initial success. However, they should always reflect rollout costs, not the higher test costs of small quantities, especially in areas of artwork, printing or postage (e.g., the test may not qualify for postal discounts).
Within the P&L, determine the biggest point of vulnerability and identify the actions that can be taken to compensate. For example, a major factor in overall profitability will be the mix between customer and prospect sales. It is helpful to do the P&L from the standpoint of actual dollars, projected dollars, and by percentage of dollars.
#4: Are We Building For the Future?
You must always work at replenishing attrition with new customers of greater or equal value. To do this, look at:
- Mix between first-time buyers and multi-buyers.
- Mix of buyers by source. This will include a look at the impact of Internet customers. Are you bringing in new customers or adding sales for existing customers or retaining customers longer?
- Sales of new first-time buyers versus sales of past first-time buyers.
A little more than a decade ago, a catalog company was sold that looked quite profitable to the new owners. What they did not realize was that the owner, in anticipation of retirement, had not done any prospect mailings for two years. Profits were up, but there had been no investment in the future. This catalog struggled for years to repair the damage done by its founder.
#5: Where Should I Concentrate Prospect Dollars?
Begin by looking at the cost per customer acquired. That is what you spent to acquire them, including list rental and total in-the-mail costs.
Next apply any conversion rate (and associated costs) that is relevant to a two-step process, such as a space ad offering free catalog. If fewer than 100% actually make a purchase, the cost per first time buyers will rise.
Then, compare these by source (e.g. by media type, promotion type, format) to maximize your most cost-effective acquisition methods.
The most valuable measurement tool you can have in this area is Lifetime Value (LTV). LTV is the net present value of expected future profits from a cohort of similar customers. It represents a best guess of the net profit a new customer will bring in over her lifetime. LTV must be developed by a statistician after a careful study. Many experts analyze at least two years worth of purchase data.
LTV is a measuring stick for what you can afford to spend in order to acquire future customers. For instance, if your LTV is $20, many companies have determined that they should spend 20 to 40 % of that, or $4.00 to $8.00, to acquire new customers. If you spend less, you are being too conservative. If you are spending more, you are risking the future profitability of the catalog. LTV may vary based on the original source of the name (e.g., space ads, rental lists, and Internet purchase).
#6: How Much Will Sales Need to Increase to Pay for a More Expensive Creative?
Determine the in-the-mail cost of the new creative in rollout quantities. Using the P&L format, calculate how much sales need to increase. If they need to go up by 50%, and creative changes have never given you more than a 15% increase, this is probably not a good way to spend test dollars.
#7: Which Test Cell Should We Rollout?
Begin by verifying that:
- There is not more than one change between the control and each test cell.
- There is no name selection bias.
- The sample size is sufficient to read significant differences between test cells.
If these issues are not resolved up-front, test dollars will be wasted and valuable time lost.
The goals of each test should be clear from the beginning. Are the objectives of the test to lower costs, increase response, or some other factor?
Construct a P&L for each test, using rollout costs. Apply different costs where appropriate. Carefully measure any element that is likely to be affected by the test (e.g., percent of orders that qualify for premiums or other special promotions).
As results come in, build a "Test Cell Summary," which is a concise summary of the key results and issues of each test. Identify results that are significantly different from each other. Indexing results can help clarify differences.
#8: Which Customers Should We Promote and When?
RFM (Recency "Frequency "Monetary Value test cells) results are often used by catalogers. And, these do give helpful insight into the performance of various customer groups. However, they will only look at one aspect of a customer's performance in ranking importance.
On the other hand, statistics-based predictive models evaluate the multiple variables and assign a specific score to the customer. Customers are then ranked from high to low and divided into groups of equal size. These groups are often deciles, or ten equal groups. A statistician is needed to assign the scores. Once the scores have been assigned to each name, the implementation of circulation decisions is much easier than with RFM.
Significant history is needed to build a statistics-based predictive model. Ideally, multiple years of historical information should be included in the model-building process. Typically, separate segmentation models are done for single-buyers versus multi-buyers, or seasonal variations.
Over time, the circulation manager can develop a plan based on the number of deciles that should receive a given catalog.
#9: Which Items are Profitable Within the Catalog?
One useful tool is the Square Inch Analysis, which measures the number of square inches dedicated to an item. The cost of the space is then deducted from the gross margin (i.e. sales minus cost of goods) that the product produced.
However, caution must be applied. If a cataloger determines that 15% of the items were unprofitable, and eliminates them from the catalog, that would increase the cost burden on the remaining items and cause some of them to be unprofitable. If the incremental cost of including an item is low, it is fine to include some items that are not profitable on a fully-loaded basis.
We have only covered a few of the many metrics available to catalogers. However, it is important to realize that the questions we ask, and the search for the answers, is first step in a meaningful measurement of success.
Never stop asking those questions.