An alternative to the book-to-bill? These new approaches can help measure long-term PCB trends more accurately.

Let me start by throwing down the gauntlet to the many protagonists of the book-to-bill ratio. Many PCB manufacturers have great faith in this technique, using it to make judgments that influence their business strategies. I would like to be contentious and suggest an alternative.

Now, I agree that the book-to-bill is one form of a market indicator, if a rather crude one. However, I don't think that the book-to-bill analysis is particularly sensitive to longterm trends. It's something of a morale booster when the market is down and possibly a dangerous harbinger of disaster when the market is up.

"Book" is the value of total orders received (or booked) during a particular time period, a month for practical purposes. Its counterpart, "bill," is the value of sales invoiced (or billed) in the same time frame. Divide the book by the bill, i.e., orders in/sales out, and you get the book-to-bill ratio. If the ratio is higher than 1.00, the book-to-bill is rising--orders are coming in faster than sales. An answer less than 1.00 means the book-to-bill is falling, and sales going faster than orders. You are running out of business.

This ratio is fine as a short-term indicator and when the short-term level of business is relatively constant. But if business is declining long-term, sales billed per month (the denominator) will be a falling number. In difficult times this could still produce a BTB figure greater than 1.00 despite the overall decreasing trend. TABLE 1 shows a simple example. As the book falls, the company slows its billings to compensate for the reduced demand. The BTB starts to fall but still remains greater than 1.00 for several months despite deteriorating business conditions. The falling bill is propping up the ratio due to the mathematics of the BTB calculation. It takes five months before the ratio dips lower than 1.00. FIGURE 1 shows a longer-term scenario. The BTB is very "busy," which tends to obscure the long-term trends. Even during much of the period of declining business, the BTB is still above unity, giving a false sense of security in deteriorating conditions.

[FIGURE 1 OMITTED]

Even if you take a longer-term view and run a book-to-bill on a three-month average basis (FIGURE 2), the trends still don't show up clearly. The curve is still busy, with plenty of short-term fluctuations and no clear message to be deduced. The disadvantage of the three-month approach is that the numbers kick in later, since we have to wait three months to get the first reading. Extending the analysis time frame appears to confer little benefit on the picture painted by the BTB.

[FIGURE 2 OMITTED]

This has led me to look for another simple ratio or index, which might signal a clearer picture. Why not get away from BTB and look at the trends in each element in a different way? If bookings and billings are rising or falling, isn't there some way we can link these two numbers?

After experimenting with a series of different options, I eventually settled on a very simple formula. I took the ratio of bookings in Month 2 compared to those in Month 1 and added this to a similar ratio for billings over the same time period. I divided by 2 to get an average and up came a figure, which I christened the "Optimism/Pessimism Index" or OPI for short.

2 Month OPI = (Bookings Month 2/Bookings Month 1 + Billings Month 2/Billings month 1) / 2

I tried this first as a simple month-on-month index and the results were encouraging. FIGURE 3 shows how the OPI plays against the sales numbers used above for the BTB. After an initial fall, the OH starts to climb gently as the sales figures start to improve. Although slowly rising, the OPI remains resolutely below 1.00 on the right-hand Y axis until the business starts to really pick up with increasing bookings and billings both moving upwards again into growth. This is a much clearer picture than the BTB in Figure 1, which is still drifting downwards when the company's fortunes start to improve.

[FIGURE 3 OMITTED]

I tried a slightly longer-term version of the OPI by spreading the figures over three months as per the following formula:

3 Month OPI = (Bookings Month 3/Bookings Month 1 + Billings Month 3/Billings month 1) / 2

This produced a similar curve, which takes three months to kick in instead of two. The two different versions of the OPI are shown in FIGURE 4. There doesn't appear to be a lot of benefit in the three-month OPI, as it seems to follow the same trend as its two-month cousin. The curve is perhaps a little smoother and a tad more pronounced in its upswings and downswings and takes a little longer to send out signals of any change in company fortunes.

[FIGURE 4 OMITTED]

Compare the different options in one common graphic, demonstrated in FIGURE 5. This somewhat chaotic chart gives a comparative picture of the BTB and OPI plotted on the same axis and each in its two-monthly and three-monthly version. The BTB tends to lie in the range 0.95 to 1.05, meandering along the gridline "index value" and oscillating as it does so. No clear trend emerges. The two-month OPI, however, starts around 0.9 and rises more vigorously to a value of about 1.05 before falling towards a final value of 0.95. The OPI curves move much more dramatically, reflecting more precisely the changing patterns of the business levels.

[FIGURE 5 OMITTED]

In one final mad moment I cooked up another index based on a combination of the OPI and the BTB. I christened this the "HAWMO" index, which stands for "How are we making out?"

HAWMO = (OPI +BTB) / 2

This final offering, shown in FIGURE 6, appears to offer the clearest picture of the market trends thanks to its linking the BTB with OPI. When HAWMO is less than 1.00, things ain't so good, but a rising value, which passes unity, should signal better times. Conversely, if HAWMO starts to decline towards or below unity then things are on the downturn again.

[FIGURE 6 OMITTED]

Figure 6 shows HAWMO in all its glory and I offer this as a new index for measuring PCB market trends. Suggestions are invited for a smarter or sillier name.

The Cumulative Plotting Approach

Years ago I attended a management training course at a local college in West London now named Brunel University. We visited several local companies, including an ice cream manufacturer, to see to see how they managed their businesses.

Selling ice cream is a seasonal business, peaking in the summer months. But making ice cream is a process business--it needs a constant, continuous level of production. Matching these two conflicting demands requires a buffer stock to be built up during the winter and spring to carry into the summer when demand reaches its maximum. Production thus remains at a fixed level all year as inventory rises and falls.

Storing ice cream requires a refrigerated cold store the size of an aircraft hangar. How did the ice cream makers calculate the stock levels needed to get them through the summer and keep factories rolling at a constant rate of production? FIGURE 7 shows the kind of annual ice cream sales pattern that the sales people experienced. How do you get from that curve to a decision about cold store size?

[FIGURE 7 OMITTED]

First, plot the sales on a cumulative basis, producing the "S" shaped sales curve in FIGURE 8. It's then easy to see the annual demand and to draw a straight line from January to December to work out the cumulative level of annual production. We now know at what production level to run the factory but we still don't know how big to build the cold store. To do this we have to shift the production curve in a northwesterly direction until it just "kisses" the sales curve around October (FIGURE 9).

[FIGURES 8-9 OMITTED]

At these levels of demand and supply, maximum stock level will be achieved in April with approximately 2.6 months of production in the cold store. This would be an extremely dangerous course to sail, as it assumes an empty cold store in October and provides no safety margin during the summer peak. If the summer turned out to be hotter than usual, then potential ice cream sales could be lost. This approach allows no latitude for error, so you'd have to build in a safety buffer.

A safer scenario is set out in FIGURE 10. The production curve has moved up further. The year begins with a higher starting inventory, leading to a higher maximum stock level by April and hence a larger cold store. The summer is covered sensibly and as fall approaches there's sufficient in hand to prevent a "stock-out" in October as the FIGURE 9 version threatens.

[FIGURE 10 OMITTED]

Finer tuning is also possible using the cumulative approach. If cold store size is a problem, a two-tier production pattern is an option. It's possible to calculate an adjusted production pattern in which the year is divided into two production segments, winter and summer. Using the cumulative curve and some mathematical juggling, a "two-phase" production pattern can be set up as in FIGURE 11 and FIGURE 12. This allows a smaller cold store to be used, as there's less need to build a spring buffer stock to see through the summer and early fall.

[FIGURES 11-12 OMITTED]

The establishment of the two-stage annual production plan in Figure 12 is only possible after using the cumulative approach set out in Figure 11. Such a plan means increasing summer production capacity--a significantly less attractive proposition from a capital investment point of view. However, the model shows how such a plan uses cumulative techniques to judge relative production levels, timing of changeovers and low stock danger points.

This process would be virtually impossible to set up by using Figure 12 alone. It's the cumulative route that allows a realistic plan to be made which sets cold store sizes and production capacities, plus fixing the points in the year at which production levels that have to be changed.

Cumulative Approach to Sales and Orders

Cumulative techniques lend themselves well to looking at sales and orders, providing valuable smoothing which helps to determine longer-term trends. The juxtaposition of the sales and order curves created by cumulative analysis can be a powerful tool in terms of order and sales planning.

Take an example of some sales and order statistics, which could arise at any typical company. These are set out in TABLE 2. After a mixed set of monthly figures, sales start to rise in April through June. The CEO, an optimist looking to grow the business rapidly, decides that the improved trend will be continued and forecasts sales to run at an average of 120 units per month for the rest of the year.

However the sales director, an older and wiser realist, takes the view that orders will roll in at an average of 100 units per month, as this is the average of the first six months' bookings.

The apparent relationship between these two points of view can be seen conventionally in FIGURE 13. It is clear from this chart that the differing opinions will lead to a final disaster as orders fail to keep up with sales, but how and when will this happen? When will the crunch come and how long does the cautious sales director have before things get really tough? Does Figure 13 give any clue except for telling everybody that somewhere ahead waits a disaster scenario?

[FIGURE 13 OMITTED]

The solution lies in the cumulative approach, which puts things clearly into perspective and highlights the danger areas. This is illustrated in FIGURE 14, showing the disaster scenario that would unfold if no remedial action were taken to avert the oncoming danger:

* The opening scene shows a backlog of 300 sales units. The horizontal distance between the sales and order curves represents the number of months of orders in hand (at that particular moment and at that particular going rate). Our curves show three months' orders in hand as the year begins.

* By the end of June (when the historical numbers come to an end) the order book shrinks to 267 units. There are now only 2.4 months' orders in hand as the rate of sales is now increasing faster than the orders are coming in.

* Fast-forward to the end of the first year and the curves show that if the situation is maintained without action, the order book will have halved and there will only be 1.3 months of orders left.

* The Armageddon point is reached the following June when the orders run out.

[FIGURE 14 OMITTED]

My point is that the cumulative graphical technique allows you to see ahead and make decisions to prevent a disaster. The changes of direction of the sales and order curves can be extrapolated to predict convergence points (or a dangerous coming together of the two curves). The horizontal distance between the curves can be used to calculate how many weeks of business there will be in the backlog at future dates.

The vertical distance between the curves allows the value of the backlog to be calculated and this data can be used to make judgements about the future trend of the business. Accumulation introduces smoothing, predictability and the facility to calculate business patterns.

In my opinion this is an intelligent and simple way to look at your business. The cumulative technique can also be adapted to other situations:

* Shipments versus production. In this case, the vertical dimension between the two curves represents inventory so that the number of weeks of stock available can be calculated.

* Raw material purchases versus production. You can control incoming material supplies and inventory levels.

Cumulative analysis a useful and practical tool, and adds another approach to the process of running your business successfully. It can help you make decisions about oncoming events in the sales arena before it's too late to change direction, assist in controlling inventories and guide you on investment planning.

TABLE 1. Despite deteriorating business conditions, the
BTB remains above 1.00 for four months

MONTH   BOOK   BILL   B-TO-B

1       120    100     1.20
2       105     95     1.11
3        95     90     1.06
4        90     88     1.02
5        85     87     0.98
6        80     85     0.94

TABLE 2. Example of sales and order figures for business

                        JAN.   FEB.   MARCH   APRIL   MAY   JUNE

Actual monthly sales    112     90     101      98    110    122
Forecast sales
Actual monthly orders    90    110     88      111     92    109
Assumed order trend

                        JULY   AUG.   SEPT.   OCT.   NOV.   DEC.

Actual monthly sales
Forecast sales           120    120     120    120    120    120
Actual monthly orders
Assumed order trend      100    100     100    100    100    100

ROGER TYLER is a principal of Appletree Associates, a consulting firm in the UK. He can be reached at [phone]; [e-mail].