Market Snapshot: Pet Food
Webinar - May 14, 2019

Market Snapshot: Pet Food

We bought bags of kibble from a variety of retailers. We tested big brands alongside small niche companies. The result? A market snapshot of thirteen pet food brands. In this quick 15 minute webinar, Scott Campbell reviews the findings.


  • which brands came close to exceeding the 10% moisture limit.
  • what some big brands do much better than small ones.
  • which brands over-pack—and which ones don’t quite meet the minimum.
  • the projected costs of inconsistency in packaging and drying.

Presenter: Scott Campbell reveals the findings of the METER Food Lab. He has visited dozens of pet food manufacturers in his quest to help them collect data and use it to consistently make great product.

Next steps:


Brad Newbold: Hello everyone and welcome to our Market Snapshot: Pet Food webinar. Today’s presentation will be fifteen minutes, followed by five minutes of Q&A. If you have a question during the webinar, type it into the questions pane. You can submit a question at any time. We will keep track of these questions submitted during this portion to answer during the Q&A. To answer our most commonly asked questions in advance, we will be sending out links to the on-demand webinar and the slides to everyone who attended today’s webinar. Those will be available a day or two following the webinar. Now I’m going to hand the microphone to Scott Campbell, today’s presenter.

Scott Campbell: Hello everyone and welcome to our discussion today. I want to start out by talking about standards, and something called the kibble balance. Although this is a pet food webinar, the kibble balance doesn’t have anything to do with pet food. There is a standard of weights and measures that everyone knows as the kilogram. What you may not know is that the kilogram is actually a cylinder of platinum and iridium that lives in a location outside of Paris and, as of the 1800s, was a standard that was created and it actually exists. You can touch it. Now there were some problems with that standard because every time people touched the cylinder, they were worried that they would either add or remove weight from it. Over the years they noticed that the weight of the kilogram, of that particular standard, had changed relative to other copies that were made of it. After many years, in fact just last year, they decided that the kilogram – the chunk of metal – needed to be retired. They replaced it with a new standard based on Planck’s constant – something that will never change – and that is implemented in something called the kibble balance, that now officially and internationally replaces the kilogram. They had one standard everybody was adhering to and replaced it with a better one. 

That’s what we’re going to talk about today – ways to improve the perspective that we take on moisture and overpack in the pet food industry, because there are lot of opportunities for improvement and we want to talk about the data that we collected on that. We deal a lot with moisture and overpack in the pet food industry and we decided to take a different approach to seeing how prevalent these problems are in the industry by simply going to the store and buying a bunch of pet food products. In this study we focused on kibble type products and we measured the moisture level, the water activity level, and the amount that the weight in the bag exceeded or did not exceed the label weight, and then we just organized and collated the data and that’s what we’re going to share with you today.

In the first place, we measured the moisture of each of these products. You can see that we had sixty or so different products, and they are listed here in terms of lowest water content to highest water content. One of the issues with moisture content is that there is no good standard for measuring it. Unlike the kilogram, there isn’t a common way that everybody can agree on to measure moisture content. Our way was to dry the product in an oven for four hours at 105°C which is an AOAC approved method. This is very slow and not good for process control, but it is pretty precise. What we found was that there was a huge variability in the moistures of the products that we tested. 

Of these sixty or so different products, you can see them listed here by first alphabet letter which is a code for the company that created that product. We listed the product name on here as well. The lowest moisture content product was from company D, and they had a lot of products that were relatively low moisture. What we found was that the lowest water content we found was a little above 4% and the highest was almost 10%. There was a huge variability.

 One thing that people will ask is, “Is it correct for companies to try to maximize moisture content in the product?”. To a certain point: yes. It is good to do this partly because, from a profitability standpoint, higher moisture content equals higher yields. From a palatability standpoint, most experts believe that pets prefer pet food that is higher in moisture. Of course the moisture content can’t be too high or you will experience mold problems or other microbial growth problems with the product. However, that’s a water activity issue, not strictly a moisture content one. 

Let’s look at what some of these data mean. Note first that even when we are comparing two different products that are the same type of product, like this duck product here, the moisture content for one of them is 5.88%, and for a different recipe we have a moisture content of 8.96%. Similarly with this salmon product, we see one that has a 5.18% moisture content and another that has a 7.89% moisture content. We didn’t see that the moisture contents of different similar recipes were anywhere close to each other. 

This is a separate graph. Here, we are translating moisture content into a graph where we are showing it on the y-axis with water activity on the x-axis. This is important because it is possible to translate between moisture content and water activity particularly for the same product. What what we’re seeing in this graph is a grouping of different dots whose colors correspond to the company that makes each product you see graphed. A close grouping of dots will suggest that the company does a good job of consistently putting out a number of different products at the right water activity or moisture content. A broad grouping will show that their process control is not that good. What we see with this company, H, is they actually have a relatively close grouping of dots on the graph but we see other companies, such as F, where the groupings are all over the place, running from a water activity of around .4 all the way to .6. Company D does have a relatively consistent performance here, but all its products are very low in moisture and water activity, all of them below .4 and below 6% moisture. 

If we look at this graph as a whole, we can tell that there is a clear relationship between activity and moisture, and that’s called a moisture sorption isotherm. One question that we often get is, “how should we consistently control moisture?”. As I mentioned before, and this gets back to the kilogram, you need to have a common standard for moisture. The best standard to use, and with us being a water activity company this will come as no surprise, is water activity. Because it is a primary measurement method, it doesn’t depend on whether you use Karl Fischer titration, rapid loss on drying, NIR, or nuclear magnetic resonance. All of these are moisture methods but they all give you a different answers. Water activity will always give you the same answer if you’re measuring the same product. And what’s more, you can see that most of the pet food products are under this .6 water activity level. In the moisture content terms, most people in the pet food industry talk about 10%, and that’s not a coincidence because 10% is about .6 water activity, and over .6 you can get mold growth in the bag. 

This is a reason why people have talked in moisture terms for so long. Empirically, people found that they had mold issues over 10% moisture, but really, that’s just because the water activity was above .6. It doesn’t matter what recipe you’re making or what product you’re making: If your water activity is below .6, you will never have any microbial growth in your product.


Not a “quick mode” reading, but a direct measure of water activity to 0.003. No calibration, no approximation.

The other point to make, though, is that precision of the moisture measurement method is very low in the dry range. This means that, looking at the graph, a small change in moisture is correlated with a very large change in water activity. What that means is that, if you choose a more precise method to measure moisture, you can know whether the changes that you’re seeing are real, and you can eliminate the measurement method as a potential source of variability. 

This is absolutely critical when you are diagnosing process issues. You need to know immediately when a statistically significant change in moisture is happening. Because a small change in moisture means a large change in water activity, the signal to noise ratio for water activity, as a way to look at moisture, is simply better than moisture. 

So let’s talk about that relationship between moisture content and water activity. That’s shown here on this graph. This is for a single product, a kibble pet food, that we created with an isotherm generator which is an instrument that we manufacture here at Meter Group. Note that on this chart you can see that a 9% moisture is correlated with a .63 water activity, and a 10% moisture is correlated with a .7 water activity. A tiny change in moisture, 1%, equals a massive change in water activity – from .63 to .7. If we take the second derivative of this moisture sorption isotherm, The sensitivity of our best what activity meter is plus or minus .003 water activity units. You can see how sensitive the math is. You can also see something very interesting that we wanted to point out, which is that the rate of change of moisture attaching itself to this kibble product, as the product becomes more saturated with water and the moisture content increases, jumps up suddenly when we reach that .7 water activity. 

If we take the second derivative of this moisture sorption isotherm, we can see that the rate of change is the highest at .7. This isn’t an accident – above this level, pet food will become unstable and its shelf-life will be significantly shortened, both by microbial growth and other degrative reactions. This is a scientific way to establish the ranges where pet food has its longest shelf life, and we do see that keeping the product below that .6 range is the range where it has its best shelf life. This is a perspective on moisture that allows you to establish when the products have their best shelf life. Additionally if you know this water activity to moisture content curve, you can simply measure the water activity and know the moisture content very precisely as well. In this way, the isotherm acts as a translator for you. Even if you prefer to converse in terms of moisture you can still do it using water activity as an independent standard.

 Lastly, we measured the overpack in the pet food that we bought, and here you can see the min, the average, and the max overfill for each brand. Overall, most pet food companies have a significant amount of overpack – weight that the consumer didn’t pay for that they got anyway, although there are examples of underpack in our data. We see that this is not particularly well controlled in the industry either. There are lots of reasons for this, a complex subject for another webinar. However, we noticed both that moisture was not particularly well controlled in the samples that we purchased, and neither was package weight. 

These both present opportunities for improvement. How big could the improvement be? We estimate that, if there is roughly $12-$15 billion dollars of kibble-based pet food sold in the United States annually, and that the average input cost for pet food is $1.50 to $2, the pet food industry could improve its profitability, overall, by somewhere between $100 and $200 million dollars simply by making sure that both the moisture level of each bag of product that they sell as well as the package weight are accurate for every batch of product. That total impact for the entire industry is somewhere between $100-$200 million dollars in profit, because those are input costs that you would avoid. We see that that overpack here is not particularly dependent on type, but some of them had a significant amount of it.

This brings us to our final recommendations. Our first recommendation is that you get a single source of truth for moisture, even if that means getting a translator. By a translator, we mean being able to use water activity as an independent standard, and then being able to express that in moisture terms. It is possible, and we have done it for many clients over the years. It helps them get better independent standard for looking at moisture across the entire business. 

Another is that we find that many companies carefully monitor in-process areas, such as the moisture content of kibble as it comes out of the dryer, but then put the product in storage bins and leave it there for a long time. There are moisture changes in the product as it sits there that secretly rob the company of profit and quality. Sometimes the product will even stick together, making it difficult to get out of the bins, causing all sorts of process issues. Areas like this in the process are not well monitored or understood. These are things that our food scientists can help with. If we know the isotherm of the product and we know the conditions that it’s stored under, we can implement systems in a real time and we can tell you and even alert you when a product has exceeded a critical limit. This can signal when a product needs to be moved the next step because it’s sitting in storage and taking on or losing moisture when it shouldn’t be. Some companies will carefully monitor processes, but then leave product and lose process effectiveness in areas that they’re not monitoring.

The third recommendation is that companies upgrade legacy equipment to allow you to nail those moisture and package weight numbers on every batch that you make. Modern technology partners like METER Group are able to work with your legacy equipment to improve its effectiveness, whether it’s a thermal process like an extruder dryer or whether it is a combo or check weigher system. Both of those are places where collecting the right data at the right time can allow you to dial in the consistency of your processes and nail those numbers. This has a significant impact both on your profitability and the quality that the consumer gets from you in every product that they buy.

BN: Awesome, thank you Scott. We have a couple minutes for questions so let’s take a couple here. There’s still time to ask questions by typing it into the questions pane, and we’ll get to as many as we can before we finish. The first question says: Can we get a copy of the data that you collected for this study?

SC: Yes, if you would like to, please just contact us and make a request afterwards. Right now we’re noting the company names by letter, but if you’re interested in which company was which letter we can share that information with you, so yes, you can get the data. 

BN: Second question: We manufacture to a certain moisture content but then the products sit unpackaged in a storage facility. Does this affect the moisture levels?

SC: Yes, this was the example I mentioned a minute ago. We find that this is really common when we talk to pet food clients and customers. Holding areas for products are a common place that products typically, depending on the ambient conditions, lose moisture. Holding location can have a great impact on the product because even if the product has been carefully monitored up to that point, as it’s put into the bag it loses moisture. You need to account for the complexity of the process and factor that holding area in to the way that the product is made and then pulled from the holding area into the packaging. 

BN: What is the actual difference between moisture content and water activity?

SC: This is probably a whole different webinar, and we geek out over this kind of stuff because we are a water activity company. Simply put, the moisture content is the amount of water by weight in a sample. The water activity is the energy status of that water. It’s essentially a measure of the energy level of water. If a product has a water activity of .6, that means that it has 60% of the energy as compared to a glass of pure water. So water activity is a measure of energy level which can tell you what can grow in the product, which tells you how fast reactions occur in the product, whereas the water content just tells you tells you how much water there is. As I said they are related, and that relationship is known if you measure it for a specific product. One is measuring how much water there is, the other is measuring the energy level of that water. 

BN: The next question asks: Can’t you prevent overfill just by upgrading your equipment?

SC: You can, and we see a lot of equipment vendors that have great new smart systems for doing that. The fact of the matter is that the equipment you already have is probably good enough for this, you simply need to have an extra layer of context and analytics put on top of it so that the settings of the instrument are correct for the product that you’re making. That is a capability that can be added to your legacy equipment instead of purchasing completely new equipment, which can be quite expensive sometimes. 

BN: All right one more question, and this one kind of follows what you mentioned previously: We have to process to below 10% moisture, so why measure water activity at all? 

SC: The biggest reason I can think of from a profitability perspective is that if you measure water activity, first of all, you can translate that into a moisture number, so if the purchaser of your product requires a moisture number or a moisture spec, they can have moisture numbers. If you’re using a method like loss on drying, and especially rapid loss on drying, which has a lot of variability associated with it, then the noise ratio is higher than the signal, meaning that you can’t really trust the numbers that you get. What that means is that you’re going to have to run at lower average moistures in order to be absolutely sure that you’re not exceeding that 10% moisture spec. If you measure water activity with a precision that’s much higher, you can have confidence and dial in that process, such that statistical variability of your moistures is significantly reduced. That gives you the confidence to operate at a little bit higher moisture number, still make sure there’s there’s no chance for microbial growth, and increase your yield. If you need to hit a particular spec for a customer, that’s fine, but make a more consistent process that allows you to have higher average moistures, and that’s going to increase your yields.

BN: Thank you Scott. That’s the last question we have time for today. If you did submit a question that we didn’t get to, we’ll have our team follow up with you by email within the next day or so. Thanks again for joining us today. We hope you enjoyed the discussion as much as we did here, and thanks again for such great questions. Scott did mention that study data are available to people who attended this webinar live. If you’d like to see that data or if you’d like more information on improving control of your products’ water activity or your package weights please reply to the follow-up email. That email will also contain links to the recording and to the slides. Stay tuned for future METER Food webinars, and have a great day.


Learn more about how to use hurdles to prevent microbial growth even at water activities higher than 0.91.