When it comes to finding new starts, circulation departments need to provide the right offer at the right price in the right places.

Seemingly simple, this statement posits three elements, but is any one of these elements actually more important, or more valuable, than the others?  

To answer this question, we recently looked to a top 50 newspaper client for which these issues were top of mind. Circulation executives were directed by management to undertake an aggressive starts campaign. The campaign was to include a low circulation price (so the second of the three elements was a given, regardless of whether it was “right” or not), in exchange for a lengthy subscription term.

But the team was given no direction by management about what delivery frequency to offer; nor were they given any instructions as to where to find those starts geographically. In other words, it was left to the circulation department to quantify and optimise these two missing variables.

Nobody would debate the notion that a news company should first focus on selling the best offer into the best zip codes. But beyond this initial obvious goal, how should a newspaper pursue marginal business?

If forced to choose, should a newspaper try to sell the best delivery frequency (a major component of “right offer”) in the worst zip codes (i.e., the “wrong places”)? Or should it try to sell the worst delivery frequency in the best zip codes? Or, from the perspective of optimising subscriber acquisition, are they equally important, i.e., does it simply not matter?


For this client, we segmented a year’s worth of paid starts by delivery frequency, because different delivery frequencies almost always deliver significantly different levels of financial performance.*

Within each delivery frequency, we then sub-segmented all paid starts by zip code and then grouped the zip codes into five performance quintiles.

This allows us to not only quantify the variance in performance among delivery frequencies, it also allows us to examine value tradeoffs between frequencies and zip codes simultaneously.

* Our preferred measure of performance is net margin surplus per copy, which is defined as weekly net margin multipled by weeks retained minus acquisition expense divided by total lifetime copies, where weekly net margin equals weekly circulation revenue plus weekly preprint revenue minus weekly newsprint and ink expense minus weekly delivery expense.


When simply evaluating performance of new starts by delivery frequency, our client experience has shown consistently that newspapers tend to over-invest in higher delivery frequencies (e.g., seven-day or “DS”), in the form of high acquisition costs per start and low introductory rates.

This tendency is understandable because newspapers are always in search of more copies from which to protect their run-of-paper (ROP) ad revenue and amortise their fixed overhead.

As such, it is not surprising that Sunday-only (“SO”) generates the most favourable performance on a per-copy basis, (although in this case, the “best” is the least negative) at US$0.10, while DS generates the worst performance at US$0.18 per copy.  

As more days are added, the average surplus per copy worsens, which is also in line with what we have observed at other newspapers. However, in this example, there is one notable exception – W3.

This aberration occurred because this circulation department did not yet have the informational tools to match the acquisition expense of each start to its projected value, and it overpaid – in the form of excessive commissions – as a result.

So even though the lifetime value of the start was greater for W3 (three-day-a-week subscriptions, usually Thursday, Saturday, and Sunday) than for SO, the acquisition cost was disproportionately worse. As such, this drove down the average net margin surplus per copy for all W3 starts.

When we layer in zip codes, the complexity and range in performance multiplies. The gap between the best delivery frequency in the best zip codes (US$0.03 per copy) versus the worst delivery frequency in the worst zip codes (US$0.22 per copy) is a substantial US$0.25 per copy.

To provide some context, consider the following “average” start with delivery frequency of four-day, an acquisition cost of US$40, average retention of 26 weeks, and a weekly net margin of US$2.

Such a start generates a lifetime net margin of US$52, which when offset against the acquisition expense yields a lifetime net margin surplus of US$12. When divided by 104 copies (i.e., 26 weeks multiplied by four copies per week), this equates to a net margin surplus of US$0.23 per copy.

A swing of US$0.25 in this metric would take this start from a respectable lifetime profit of US$12 to a loss of US$14. This means that the start, which cost US$40 to acquire in the first place, would no longer generate a return adequate to offset the investment.

A range of this magnitude is also consistent with what we observe at other newspapers, and it speaks to the underlying importance of the Matching Principle, which proposes allocating acquisition investment resources in proportion to the best performing segments. Clearly, the best performing delivery frequency in the best performing zip performs at a rate much better than average.

However, this range in performance also exists within a given delivery frequency, and this range is not constant – the range of the best and worst performing zip codes narrows as the delivery frequency rises.

As seen in the chart, the range from the top 20% (US$0.03 per copy) to the bottom 20% (US$0.26 per copy) is a substantial US$0.29 per copy for SO starts, falling to a range of only US$0.08 per copy for DS starts.

Because SO starts have the biggest range in net margin surplus per copy, optimal targeting – although always important for every start – is most important for this delivery frequency, since a targeting error could deliver the biggest drop in performance, on both an absolute and relative basis.

Why is this the case? In part, it is driven by the very nature of SO starts, which is that the number of copies delivered per week is the smallest of any of the delivery frequencies.

When combined with the fact that the acquisition cost per start is not proportionally lower for SO starts (and often should be higher, per the Matching Principle, since SO starts typically deliver higher performance and thus warrant higher acquisition investment), the need to acquire the right kind of SO start has never been greater.

Thus, which is the lesser of two evils: the pursuit of the worst zip codes with the best delivery frequencies, or the best zip codes within the worst delivery frequencies? The answer depends on your goal and your tolerance for risk.

In this example, pursuit of the top 20% of all zip codes with the worst delivery frequency (DS) generates a loss of US$0.14 per copy, versus a more substantial loss of US$0.26 if the pursuit is switched to the worst 20% of all zip codes with the best delivery frequency (SO).

Thus, if managing the downside is the primary objective, DS is the preferred frequency. However, if pursuit of the maximum upside (even with greater risk) is the primary objective, then SO is the preferred frequency.

But this is only true if the starts are being found in the best zip codes. If this targeting effectiveness cannot be assured, then it will likely be better to switch to a delivery frequency with less severe downside. In this example, the worst case scenario is W4, W5, or DS is not nearly as severe as it is in SO.

Said differently, for this client, SO represents the classic definition of a “high risk-high return” strategy choice, whereas DS represents a “lower risk-lower return” strategy.


While it is always the hope of every circulation department that it can focus as many resources as possible on the best delivery frequencies within the best zip codes, pressure to hit circulation goals forces departments to pursue business that is marginal in quality.

Quantifying the tradeoffs between the best and worst delivery frequencies and zip codes can provide a strong starting point for a circulation department’s optimisation goal.