Financial Management: Costs

Post 4 in the series on management.

In the previous two posts in the series I showed how to decompose revenue growth into its constituent parts (volume, mix, and pricing), and how companies determine what prices to set for their products and services. In this post I will discuss how companies think about the cost side of things. I will be focusing primarily on COGS (Cost of Goods Sold), but will also briefly discuss cost allocations. In particular, I will cover the following:

  • Main components of COGS, and some ideas for savings for each of them;
  • The difference between gross margin and gross profit, and how to think about profitability;
  • Direct vs allocated costs.

I will not discuss marketing, important though it is, because I think there are much better resources out there. I strongly recommend the book How Brands Grow by Byron Sharp on what constitutes good marketing.

This post is mostly geared towards businesses that sell physical goods; however, the discussion on margin vs profit, and cost allocation, is applicable for service providers too.

COGS refers to the costs of manufacturing a product or delivering a service; usually, the costs of storing or transporting a product are also included in COGS.

The COGS for physical goods can be broken down into the following four buckets:

  • Raw materials: the materials required for producing the product the consumer will use. For example, this includes steel for cars, chemicals for shampoos, plastic for razor handles etc.
  • Packaging materials: the materials for producing the product’s outer packaging. For example, plastic for shampoo bottles, cardboard for shelf-ready units etc.
  • Manufacturing expenses: the cost of converting the raw and packaging materials into the finished product. This typically includes the salaries of factory workers, the utility costs of running the factories, repair & maintenance, and the depreciation of the buildings and equipment used in production.
  • Transporation and Warehousing: the cost of storing and transporting products.

The COGS for delivering a service can be harder to define, so this section of the post focuses mainly on physical goods.

Obviously, companies aim to reduce costs as much as possible, everything else being equal (everything else being product quality, brand image etc). Here follow a few ideas that can help optimise costs.

a) Scale

In general, companies tend to treat COGS as variable costs for the purposes of price-setting exercises. This is not entirely true. The manufacturing costs include many costs that are fixed (e.g. depreciation) or fixed to an extent (e.g.workers’ salaries; I say to an extent because if production increases significantly, you may have to hire more workers (or let some go, if volume drops)). In addition, scale can lead to savings even for costs that are otherwise variable (such as raw and pack materials); this is because buyers of large quantities can negotiate better prices.

Hence, one way to reduce costs (well, cost/product, not absolute cost) is to increase scale. There is no easy way to do this given the intense competition in free markets, but I want to point out one avenue that is often overlooked: producing private label products. You know when you go to a supermarket, and see things like Tesco own brand? These are not manufactured by Tesco, but by third parties — often by the same companies that produce the branded products that directly compete with private labels. Why would companies want to produce goods on behalf of a retailer, if these goods will be sold at a lower price (and are therefore likely to be less profitable)? One reason is that companies know supermarkets will stock private label products, and they would rather be the ones to provide them; it is better to cannibalise your own sales, than have someone else do it. The second is scale: if private label products are built using the same materials as the own brand ones, and at the same factories, then the cost/product for all products decreases.

(There is an interesting article on this here.)

b) Raw and packaging materials

I think most raw and pack material savings are delivered thanks to the work of engineers and scientists who come up with better formulas, or cheaper materials that are just as good . But there are a few things commercial managers can also do to lower costs:

Improve the product design. You’d think this is hard to do, because most companies — certainly large, sophisticated ones — would already have optimised product design. But there are two reasons this is not the case:

First, a lot of people look at a product without really seeing it. They inherit a range of products to manage, and take it for granted that the products have to look they way they do (unless they do a complete restage of the brand).

Second, some managers believe that costly design features are necessary to maintain the brand’s image. It is absolutely true that premium products need to look premium; but any product, no matter how ornate, could always be made more expensive. It is unlikely that a product’s current design is at the exact optimal point that balances cost and design, so managers should look for choices to reduce cost. Finance managers in particular need to highlight all such choices, and share them with the marketing / consumer research departments, who can then make a final call. For instance, suppose a company could change a bottle cap to be made of cheaper plastic, saving $1,000,000 per year. The question the company’s managers need to answer is, is the existing cap really driving $1,000,000 of value per year? Managers will have to use their judgement to make a call here, but it’s easier to make the call if they have the numbers.

Here is how I recommend managers review their product design in a structured manner:

  1. Create a list of all your products;
  2. For each product, list all its raw and packaging materials;
  3. For each material/product combination, list the dosage of the material required to produce one product unit, and the cost of that dosage;
  4. Ask the engineering team to help you classify each material; the classification system really depends on the kind of product you are selling, but at minimum it should classify materials first into raw and packaging, and second into functional and cosmetic materials wherever possible (packaging materials can also be further classified into primary packaging (bottles) and outer packaging (cardboard etc));
  5. Consolidate all this data.
  6. Ideally, do this for competitive products: just go out and buy some, and send them to your engineers to break down. At the very least it should be easy to determine the cost of the packaging materials.

This is what this could look like:

Figure 1: Product breakdown

Once you do this, you will know the exact cost breakdown of each product, and can use this for analysis purposes. Leave the analysis of functional raw materials to the experts (though do challenge experts too, if you see things such as one product variant having significantly more expensive functional ingredients — or a higher number of ingredients), but do look at the following:

  1. Spend on cosmetic and packaging materials; these are the ones on which you need to challenge your marketing department, who need to justify these costs. (To do this, take the cost per unit for cosmetic materials, and multiply it by the number of units sold in a year.) And do challenge everything: how much fragrance does your product really need? Do small designs on the product actually affect consumer behaviour? Do you really need that much cardboard around the product? Are stickers necessary? (Really do look at everything, no matter how small: I once identified a few hundred thousand dollars worth of savings by eliminating a plastic band used to bind two bottles together; apparently, one retailer had requested this binding, a request that someone granted without pushing back. Perhaps the retailer really needed this binding, but one should still ask: how much is this binding worth to the retailer? If the cost is $200,000, could we eliminate the band, and give the retailer a $50k rebate instead?)
  2. Spend per material type per product unit. This will help you identify product design choices that you were not aware were driving cost increases. For example, if you compare two different shampoo variants you may notice that one has more expensive bottle caps, or more expensive fragrance. In addition, it will help you identify costs that scale with size when, arguably, they shouldn’t. For example, I once ran this exercise and noticed that larger laundry liquid bottles had more expensive labels. This was because both product sizes had labels that covered the entire bottle — even though the larger bottle could have used the smaller labels. Perhaps the product wouldn’t look as nice in a large bottle with a smaller label — but perhaps it wouldn’t matter to consumers; either way it’s useful to know the cost of the decision to use larger labels.
  3. Look at spend per material per product volume. This will help you identify costs that scale by more than they should. In the aforementioned example of label size, if you looked at label cost per ml of product, you wouldn’t notice anything funny, because the size of the labels of the bigger bottles was larger in proportion to the bottle size increase. But when I was running the analysis, I noticed something else that was strange: there was another product, sold in plastic bags. For that product, the cost of plastic per kg of product was much higher for the larger size. It turned out this was because bigger bags had extra plastic at the end, with holes cut through that could serve as handles. Again, perhaps there is good reason for this — but it is useful to know it, so you can challenge it.

Reduce the number of products you sell. Why does reducing the number of products your company makes reduce cost? Because the manufacturing process inevitably wastes some raw and packaging materials, and the more products you make, the more materials are likely to go to waste.

For example, if you are making shampoo, at some point in the manufacturing process you are going to have to mix all the raw ingredients together in a tank. If you only make one shampoo variant, you don’t really have an issue: as soon as the tank is empty, you refill it to make a new batch. But if you make two variants, then once you finish mixing a batch of the first variant, you have to clean the tank. Some of the first variant mixture will have stuck to the tank’s walls, and will therefore be thrown out, instead of put into bottles to sell to consumers. This doesn’t sound that bad — how much product goes to waste this way? 0.5%? 1% max? Perhaps — but if you are supplying the Chinese market, 1% waste can equate to hundreds of thousands, if not millions of dollars.

(You could, of course, have two tanks instead of one; but then you incur higher depreciation costs for the second tank.)

One way to analyse your product portfolio is to create a chart with volume (i.e. quantity sold in a year) on the x-axis, and profit/unit on the y-axis; to make the chart more readable, I suggest you plot the volume and profitability percentiles rather than absolute values. This is what this could look like (and here’s a model you can use to create this kind of chart):

Figure 2: Product Portfolio Analysis

Products in the top right quadrant are fine: they have high volumes and high profit/unit. The ones on the bottom right need fixing: they have high volumes, but low profitability — you should try to find cost savings (or increase the price). The ones at the top left are profitable, but sold in low quantities — you need to push them harder: increase marketing, run promotions, etc. The ones at the bottom right are not profitable, nor sold in high quantities: these you can ‘kill’ (i.e. stop producing and selling).

Unfortunately I cannot go into too much detail on these, because I actually do not know whether the savings identification process I used for these costs at P&G is proprietary or not. Also, though I have experience in delivering manufacturing savings (I used to work in a factory in China), I have very little experience in supply chain management.

So, all I can share is the following (regrettably vague) advice:

  • Again, challenge every decision that leads to higher costs: people often make such decisions without knowing the facts. For example, I remember once my finance director found out we were renting a hugely expensive warehouse in the centre of a major metropolitan city. She asked why we couldn’t rent one in the suburbs instead, and was told that clients required one-day deliveries. But first, there was no evidence of such a request from clients (this was not in a B2C context, where one-day deliveries are increasingly the norm), and second, the cost of renting the warehouse was so high that it was hardly justifiable anyway.
  • Identify time periods when nothing is being produced, and work to eliminate them; such periords include change-overs (when you shut down production to switch from making one variant to making another), repair and maintenance, etc. (Reducing the number of products you make leads to manufacturing savings as well as raw and packing material savings because it leads to fewer changeovers.)
  • Benchmark: if your company has multiple production facilities, run benchmarking analyses against each other. Do this in detail: one plant may have a lower cost structure overall, but it may be more efficient in its electricity consumption. Benchmark vs competitors, too: it’s not always easy to do this, but there are things you can do that are perfectly legal — for example, you can use google earth to check how large their facilities are; you can read their annual reports or look at their employees’ LinkedIn profiles to see what kind of savings projects they are working on; you can buy their products and look at their batch numbers to try to understand how many units they produce in a single run, and how many products are rejected for quality defects; you can see where the products sold in each market are manufactured; etc.
  • Be creative: I know this is not a good piece of advice, because it is not concrete, but do challenge yourself to think outside the box. There is a nice anecdote to illustrate this: there was once a company whose automated production line would sometimes fail to fill in boxes, so that clients would receive empty packages. The company hired an expensive engineering consultancy that installed a conveyor belt with a weight sensor; whenever an empty box passed over the conveyor belt, an alarm would ring, and a factory employee would go and remove the box from the line. The first week, the system worked as expected. The second week though, the system did not record any empty boxes; the company’s managers thoughts the machine had malfunctioned, but they checked, and clients had not received any empty boxes either. So the managers visited the factory to find out what happened. Why did the conveyor belt stop recording empty boxes? “Oh yes,” the factory manager responded, “the alarm that rung whenever an empty boxed passed over the belt was too loud and annoying, so we placed a fan just before the conveyor belt; the wind from the fan blows empty boxes off the production line before they get to the belt”. Think in the same vain: find creative (and cheap!) solutions to your problems.

Gross profit is revenue less COGS; gross margin is gross profit divided by revenue, and is typically expressed as a percentage. Everything else being equal, the higher the gross margin, the more profitable the product; because of this, because gross margin is easy to wrap one’s head around, and because it is such a standard metric in business, many managers use gross margin to make decisions such as what products to produce or stock.

But everything else is almost never equal, and therefore focusing too much on gross margin often leads to mistakes; one of the most consistent themes in my career has been trying to move discussions away from margins to aboslute $ figures: after all, neither employees nor shareholders are paid in percentage points. In other words, what I often found myself trying to do is get people to stop defining ‘profitability’ as ‘gross margin’, and to start defining it as ‘gross profit per unit’, or, even better yet, gross profit, period; after all, as we saw in the previous post, sometimes lowering the price (and hence often the profit/unit) leads to higher profit overall.

Here’s an easy way to demonstrate why making decisions using gross margings can lead you astray: would you rather own a company that sells $1,000 bicycles with a 20% gross margin, or one that sells $100 bicycles with a 50% gross margin? Assuming you’d mange to sell the same number of bikes with either company, you should obviously pick the first one, even though the gross margin is less than half that of the second company: 20% * $1,000 is $200 — in other words, your profit per bike in the first company exceeds the revenue per bike in the second!

This is an exteme example to make a point; even the thickest manager would get this immediately. In real life, when looking at similar products, gross margin and gross profit per unit are closely linked, so it is in fact often the case that the product with the highest gross margin will also have the highest gross profit/unit. Still, I’ve come across many situations where it made sense to lower gross margin to either increase absolute profit/unit, or to increase total profit.

(Another danger is overlooking the risk of a cheaper product with a similar gross margin cannibalising the sales of a more expensive product: if you are currently selling a product for $100 and a 20% gross margin, and launch a new product with the same gross margin but at a $90 price point, you risk losing money if the sales of the new product are not incremental, but cannibalise the sales of the existing one; yet sometimes, people will say ‘ah but the two products have the same gross margin, so there is no profit risk’.)

All this sounds very simple — elementary even. So why do people (really qualified people) get this so wrong, so often? I think there are four main reasons:

  • First, some people have just had it drilled into them that they need to focus on gross margin. It’s hard to break (or even question) a habit that has been instilled since your first day at work.
  • Second, people sometimes have the wrong incentives: retailers, for example, often measure gross margin instead of absolute profit, or even absolute profit/unit (I imagine this is partly to ensure consistency across different products). As a result, there were many cases where my colleagues and I struggled to convince retailers to implement plans that would grow both their and our profit, because the retailers were too focused on growing their margin instead.
  • Third, there is often an anchoring bias: if you are in charge of a product with a 50% gross margin, and someone suggests reducing this to 25% to increase profit, you will think they are crazy — even if it makes perfect sense and absolute profit goes up.
  • Finally, sometimes people do not have visibility on other costs, such as marketing or sales costs. In such cases, people are often told ‘we need a gross margin of at least x% to cover other costs’. But there are a few things wrong with this: a) again it would be better to switch the discussion to $/unit rather than gross margin for the reasons discussed above, and b) unless the other costs are also variable (and variable not in the sense that if you change them, sales will change, but the other way around — i.e. that if sales change, these costs would also change), then the proposition ‘we need x% margin to be profitable’ is wrong.

So, if you’re going to take one thing away from this post, it is this: next time someone objects to a decision because it would ‘dilute margins’, challenge them to prove that absolute profit would actually decline. Margin is not profit.

Imagine you run a company that has several divisions or brands. You ask your CFO to prepare a P&L (Profit & Loss) statement for each division, and she comes back with this:

Figure 3: P&L by division

You notice that division 2 is unprofitable, and order that it be shut down. You are surrounded by yes-men who leap to action and execute your decision; you ask for an updated P&L, excluding the shut down, expecting to see an uptick in profit. Instead, you are shown this:

Figure 4: Updated P&L

You managed to take a healthy company and turn it loss-making; you are promptly fired. Your alma mater no longer lists you among its distinguished alumni. What went wrong?

The issue here is that the company incurs some costs that are not directly driven by any of the three divisions. Your CFO allocated these costs to each division based on the division’s revenue, and it is the allocation of these costs that made division 2 appear unprofitable. Shutting down division 2 eliminated its direct costs, but the company’s overheard costs were not reduced; in the second table, they were re-allocated to the remaining two divisions.

This is what the business looks like with the overheard costs carved out:

Figure 5: division P&L + overheads

As this table makes clear, Division 2 is actually profitable; shuting it down eliminates the $140 of profit it contributes.

Over the longer term, eliminating a division can lead to lower overhead costs too: for example, a company might move to a new, smaller building, thus saving on rent; it might shrink the size of its management team; it might end up spending less on managers’ travel costs; and so on. But these benefits rarely manifest immediately.

Of course, classifying costs into direct and overheads is not as simple as it may sound. For example, if you have a marketeer who spends 80% of their time on brand A, and 20% on brand B, do you consider their salary an overhead cost? Or allocate it to brand A? Or split it 80/20 between the two brands?

Exactly how you should allocate costs depends on the purpose of the exercise. If you want a rough idea of how profitable each business unit is, I suggest allocating costs based on effort rather than something simpler such as revenue. The general principle here is to understand the most fundamental driver or function of each cost, and use that to allocate it. For example:

  • The primary function of machinery used in production is to make goods. If one machine produces products for different brands, instead of allocating its costs (repair and maintenance, depreciation etc) based on something like brand revenue (or even brand COGS), you should look at how many units of each brand the machine produces, and split the costs using that.
  • The primary driver for office space (well, other than location…) is size, which in turn depends on number of employees. So office rent should be allocated to brands / geographies etc based on the number of employees working on each brand / market etc.
  • For employee salaries, you can ask people to declare on which divisions/geographies they spend their time, and allocate costs accordingly.

This should give you a fairly accurate idea of how profitable different parts of your company are. But if you are doing this exercise because you are planning to spin off or shut down a division, you need to be more precise. For example, if two divisions share factory equipment, the cost for that equipment won’t disappear if a division is spun off; on the contrary, the divested division will incur higher costs, as it will probably need to acquire a machine for its exclusive use. Similarly, the 20% of an employee’s cost (based on that employee’s effort) that is allocated to the division about to be shut down will not disappear either.

This concludes the costs discussion. One more substantive post to go, on modelling best practices and some other bits and pieces (how to do sensitivity analyses etc), and maybe one final post with a broader discussion about big data, automation etc.

I work at Monzo. Ex Google, P&G