Velocity questions have dominated the CPG Data Tip Sheet inbox this month, with a number of readers posting comments or writing with questions about this important measure. So we decided it’s time to turn our attention to what is, after all, the third most important measure in your database.

Why is velocity the #3 measure? Well, sales is #1 because that’s your bottom line. You wouldn’t buy data if it didn’t include sales. And distribution ranks #2 because you can’t have sales without distribution. Velocity (or “Sales Rate”) ranks third in important because it captures everything other than distribution.

Velocity tells you how well your product sells *when it’s available to consumers on the shelf*. When you combine velocity and distribution, you get retail sales.

In mathematical terms (and why not do some math? This is data class after all!):

I’m going to illustrate this concept with the simplest of examples: unit sales in a set number of stores. We don’t recommend using this simplified measure (Units per Store) for real-world analysis, but it’s great for getting a handle on the concept. (I’ll talk about which measures we do recommend later in this post.)

Say there are two brands, both with sales of 3,000 units. But the forces driving their sales differ. Product A has low distribution and high velocity. Product B has excellent distribution but weak velocity, relative to Product A.

The difference between these measures influences the approach you would take to increase sales. Distribution is ultimately under the retailer’s control. So for Product A, you would make a case with the retailer for greater distribution to increase sales. Velocity can reflect a variety of factors, many of which are under the manufacturer’s control. So for Product B, your next step would be to dig in and assess the factors that impact velocity, such as price, promotion, variety or competitive factors, and then prioritize your marketing programs based on your findings.

**Choosing the Right Velocity Measure**

If you rearrange the sales equation, you can better understand the basic idea behind the many velocity measures in your Nielsen/IRI/SPINS database:

Every velocity measure contains sales in the numerator and distribution in the denominator. The measure you choose will depend on whether you’re comparing* across* different markets/retailers or are looking *within* a single market/retailer.

The two major velocity measures are:

- Sales per Point of Distribution (SPPD)
- Sales per Million (Sales per $MM ACV).

Companies occasionally use a third measure, Sales per Store. This is the measure I used in my simple example above. We don’t recommend this measure because it it doesn’t take into account differences in store sizes, which can can create biases in the velocity estimates.

**Sales per Point of Distribution (SPPD) **

Looking at velocity as Sales per Point of Distribution (SPPD) is the simplest, most intuitive choice for analyzing sales within a single market or single retailer. Also sometimes called “Sales per Point,” this measure is an improvement over Sales per Store because it incorporates store size by using an ACV weighted distribution measure in the denominator. Like all velocity measures, SPPD can be expressed in dollars, units, or volume per point of distribution.

SPPD is commonly seen in ranking reports at the item level for an individual market, as in this example:

By breaking sales into distribution and velocity, we have divided these five products into two groups. One group has strong distribution and poor velocity (Product 1, 2, and 4) and the other group has poor distribution and strong velocity (Products 3 and 5). This is a great starting point for further investigation.

SPPD works at the retailer, market, channel or total U.S. level as long as there is only one market in your analysis. Don’t use SPPD to compare across markets, channels, retailers or to compare a retailer to its remaining market (ROM).

For velocity to matter, you need to have variation in distribution. When distribution is near 100% within a market, velocity and sales rankings will be almost the same, so you can limit your focus to sales. For example, in the table above, Products 1, 2, and 4 all have similar distribution. When we compare these three products, analyzing dollars and velocity will lead us to the same conclusion.

**Sales Per Million (Sales Per $MM ACV)**

When you’re comparing *across* markets, SPPD is no longer a helpful measure. Instead, the best velocity measure for comparing across markets is Sales per Million. Sales per Million should be used when comparing a retailer to its remaining market, comparing across metro markets or census regions, and comparing one channel to another.

Why doesn’t SPPD work when you are comparing across markets? Because a point of distribution in a big market will yield more sales than in a small market. For example, distribution in 75% of New York will generate way more sales than 75% distribution in Des Moines. When you’re looking across markets, not only can the % ACV differ but the actual ACV varies too. So you need a velocity measure that reflects actual ACV, not % ACV. That’s what Sales per Million does.

Sales per Million is short for “sales per million dollars of market ACV,” and is calculated with this formula:

Another way to think about it: For every million dollars of total market sales, X amount of my product is sold.

As you can imagine, market ACV is a really big number, which is why it’s expressed in millions. It just makes it easier to work with.

As with the other velocity measures, Sales per Million can be expressed in units, dollars, or volume.

To illustrate, let’s return to our earlier example, but here we want to compare Miami to Jacksonville. Miami is three times the size of Jacksonville. The raw sales numbers reflect that population difference and the velocity numbers do too. For example, look at Product 1: Dollars per Point of Distribution is much lower in Jacksonville than Miami.

So does the lower velocity in Jacksonville reflect weaker consumer demand or just a smaller playing field? To find out, we turn to Dollars per Million, which adjusts for the difference in market size by putting market ACV in the denominator. Now we can see that, controlling for market size, only Product 3 has a better sales rate in Miami. The rest of the items sell better in Jacksonville.

Using Sales per Million is key when comparing velocity across markets. You can also use Sales per Million within a market, instead of SPPD. Some people prefer to use SPPD whenever they can because they find it easier to understand and/or communicate. Other people prefer to use Sales per Million in all their analyses so they are consistent in their velocity measure.

That’s it for today’s velocity class! Here are the top four points to remember:

- The first step in understanding sales is to separate distribution and velocity.
- Velocity is always some measure of sales divided by some measure of distribution.
- When looking within one market (not across markets), you can use either Sales per Point of Distribution (SPPD) or Sales per Million (Sales per $MM ACV).
- When looking across markets, use Sales per Million. Do not use SPPD.

Need help incorporating velocity into your analysis or presenting your sales rate findings to a non-expert audience? Contact us to learn more about our consulting services. We provide analytic support, hands-on analysis, and syndicated data training for individuals and groups.

Did you find this article useful? Subscribe to CPG Data Tip Sheet to get future posts delivered to your email in-box. We publish articles every 2-3 weeks. We will not share your email address with anyone.

*Note: Data examples in this post were edited to eliminate reader confusion caused by spreadsheet rounding. SM 2/3/17 *

Hi – I have both units per point of ACV and dollars per point of ACV in my database. Does it matter which one I use? If I used dollars and my competitor is priced higher, would that show higher velocity for them?

Your choice of velocity measure (dollars or units) follows the same logic as your choice to total sales measure (dollars or units).

When creating a selling story for a retailer, dollars per point of ACV would likely be the better measure to use. Dollar measures are typically the most compelling for that audience.

Manufacturers often prefer to look at unit or volume measures, since they don’t control the retail price. However, I think it’s still helpful to look at dollar measures as well, since it will factor in those competitive price differences and give another view of the bottom line. That same logic would apply to dollar velocity. If you look at both unit sales and dollar sales, you would probably want to look at both velocity measures too. If you never look at dollar sales and you always analyze units, then you want to focus on unit velocity. You would generally line up your velocity measure choice with your total sales measure choice.

If your competitor is priced higher, their dollar velocity might or might not be higher. A higher price does not guarantee higher dollar velocity, just as a higher price doesn’t guarantee higher dollar sales.

Is velocity additive?

No, velocity is not additive. To understand why, it helps to remember that

velocity=sales rate. Rates are not additive – to measure the rate of speed of a group of cars, you wouldn’t add their individual speeds together.To combine velocity measures across products, markets, time, etc. you can either take an average or, if you have the underlying data available, re-calculate the measure for the group. Either tactic should be done with care!

Greeting Sally,

I hope this post finds you well.

You have stated the formula as:

Sales Per Million = Sales / (% ACV * ( Market ACV / 1,000,000))

Can you confirm that %ACV is a required in the denominator for this calculation?

The Nielsen training material I have received states that:

$ Sales Per Million = $ Sales / (Market ACV / 1,000,000)

In other words, it does not include the %ACV in the denominator.

Can you clarify which is correct?

Thanks in advance.

Hi Nestor,

Here’s my conceptual “what does this measure mean” answer: in calculating Sales per Million for a product, the level of distribution for the product

mustbe reflected somewhere in the denominator. The whole point of the measure is to control for distribution.Why does your training material formula not seem to reflect this? The most likely explanation is a different meaning for the term “Market ACV”. When I used that term, I meant total ACV for the entire market across all products. Nielsen’s “Market ACV” may be product specific and reflect the total ACV for stores

that carry that product. If that’s the case, in your Nielsen data, there would be different values for “Market ACV” for each product. If they are using “Market ACV” in the sense that I was, the value for that measure would be the same for every product in the market.Does that make sense?

Best,

Sally

Hi Sally,

Yes, that makes perfect sense. In fact, that explanation helped me derived the following algebraically.

Assume %ACV = (Σ ACV of Stores Selling) / Total Market ACV

If…

Sales Per Million = Sales / (%ACV * ( Total Market ACV / 1,000,000))

By simple substitution…

Sales Per Million = Sales / ( (Σ ACV of Stores Selling) / Total Market ACV ) * ( Total Market ACV / 1,000,000))

Which means…

Sales Per Million = Sales / ( (Σ ACV of Stores Selling) / 1,000,000 )

Which is exactly what you are saying. In the Nielsen training material, the denominator is a bit ambiguous.

Thanks for the clarification on that one.

I think I have another question/comment about this topic. I need to finish working through it on my end before I post.

Warm Regards,

Nestor

Hi Sally,

I had asked you a question offline about some discrepancies I was encountering while trying to back into the Total Market ACV using Sales Per Million. Your feedback enlightened me, so I wanted to share it with the rest of your readership.

Sally wrote:

“The way I back into total ACV is to divide dollars by dollars per million. To get an accurate number, you have to use item, brand, or category totals that have 100% distribution. I use weekly numbers to do the calculation because 100% distribution weekly is a tougher standard to meet than 100% distribution in a 4 week period. I find that if I do the calculation with different items that all meet this 100% distribution criteria, the ACV number are *very* similar. But still not exactly the same. Within 1% though.”

Thanks for all of your help.

-Nestor

When looking at the brand level for velocity across accout, can I use the SSPD measure? I am looking to find out how accounts differ in velocity with a line of new items we launched last year.

Dear Kimberly,

SSPD is not a good measure to use when comparing across accounts. The sales potential of a point of distribution can vary dramatically across accounts. For example, regardless of the product, a point of distribution is going to deliver more sales at Walmart than Rite Aid, right?

To compare across accounts, the best measure is sales per million dollars of ACV.

You can see why by looking at the chart in the article that compares these two measures for Miami vs. Jacksonville. The same issues will hold true when comparing across accounts.

Now, if the accounts you are comparing happen to have VERY similar levels of ACV, then SPPD and Sales per Million might not look that different. But that’s a rare situation.

Thanks a lot for your clear explanations and useful insight.

In terms of comparing the performance of different variants within a brand, isn’t it the same to use the Share in Handler (or Share in Sellers) custom measure? What more does Velocity has to offer in this case?

If all the items within the retailer (handler, seller) have the same distribution, then yes you can just use share within the retailer. But…that is a big if! It is possible and fairly common for stores within a chain to sell slightly different items of a given brand. Also, keep in mind that an item must actually scan to be considered in distribution so the distribution is often slightly different by item, especially in categories that are typically purchased only a few times a year. Velocity controls for any differences in distribution, even within a chain. If the distribution across items really is the same, then share within the chain will be fine to use.

Hi Sally,

When calculating SPPD, is it ok to use any time frame such as 52wk, 24wk…etc? Also, is it ok to us Avg % ACV? Besides Max % ACV I don’t see any other ACV measure in my SPINS dataset.

Thank you,

Cynthia

Dear Cynthia,

I generally like to use 12 or 13 weeks for SPPD. Conditions can change considerably over a longer period of time. I will also use 4 weeks on occasion though that can be be judged as “too short” just as 52 weeks is “too long”. IMO 12 or 13 weeks tends to be “just right” for making an argument. The exception would be for a very new product where you don’t have a full 12 or 13 weeks of distribution.

Average % ACV is fine. You can also use Max % ACV if you think that’s a fairer measure of a product’s true distribution (if it’s a slower moving product category).

Hi,

I am unable to understand how SPPD values are calculated. In the example you mentioned that Sales for Product A as 616,768 and %ACV as 69%.

Based on the formula provided, SPPD is Sales/%ACV which results in the value 8939. However, in your example the value provided is 8889.

Can you please help me understand where I’m going wrong.

Hi there, You are understanding/doing the calculation correctly. That slight difference is due to rounding – I believe my %ACV numbers originally had one decimal place and I rounded them when I created this graphic. I apologize for the confusion and will make a note to adjust the chart in my post. Best, Sally

I had this same question about the calculation discrepancy, I was going crazy, thank you!

I’ve now edited the charts to resolve this confusion. Thanks again for pointing it out!

Hi Sally

I don’t understand ACV, what does this mean?

Thanks alot

We have an article that should help you understand All About ACV. If you have specific questions after that, please let us know.

We’re trying to set goals for our new retail product (ZestTea.com). We know about how many units we need to sell per month between our SKUs to retain shelf placement (about 5.1 units per month per SKU for tea). After how many months should we aim to get the product up to acceptable sales velocity?

Also, do you have any rules of thumb around how much as a percentage of annual revenue (sales to distributors) should be spent on marketing (promos + demos), to get to full velocity? We are aiming to retain about 40% margins on sales to distributors before promos and OIs.

Good questions but not something I have insight into. I would think the “how long do we have” question would vary by retailer. Sorry not be of more help on this one.

Can you help me better understand Base Velocity.

If Base Velocity has declined vs YA, does that just mean that items are just not moving as fast as they did YA? (when not on promotion)

For example:

Looking at Total Clorox Wipes (4 SKU’s in distribution)

Total Points of distribution (ACV and # items) remained the same vs YA, however Base Sales declined like $20,000 vs YA

Can you better explain how Base Velocity impacts such a large decline if ACV or # items didn’t change?

Base velocity could be impacted by many factors: everyday price changes, attractiveness of the options versus competition, consumer promotion like coupons, changes in level of advertising, changes in what competitors are doing, new products that are stealing sales from you, changes in consumer tastes. So there is a lot there to investigate and much of it can’t be examined through syndicated data. But looking at everyday price and what’s going on with your competitors would be a good place to start.

Hi Sally, Thanks for this article super helpful. Question if your $/SPP (Sales/%ACV) YOY is positive and your $/TDP YOY is negative. What does this suggest/mean?

Hi Amit, Are you asking this question about data at the brand level? If so, then the pattern you suggest would occur when you have added varieties or sizes or whatever at the UPC level (so TDP goes up but % ACV for the brand does not) and those new items deliver a a sales rate that is lower than the brand average. This situation would be a classic example of the value of assessing your brand health based on TDP versus %ACV.