Total Distribution Points: Master of All Distribution Measures

TDP chart

We have already discussed 2 key measures relating to distribution:  % ACV Distribution and Average Items Carried.  For product aggregations above the item level, % ACV Distribution provides a measure of distribution breadth and Average Items Carries covers depth. But what if you want to have one measure that summarizes distribution, taking into account both breadth and depth of distribution?  Total Distribution Points (TDP) is what you’re looking for!  (Note that it is sometimes called Total Points of Distribution, or TPD.  I will refer to it as TDP throughout this post.)

Luckily, TDP is on most IRI/Nielsen databases, although it is very easy to calculate if necessary.  TDP is the sum of the % ACV Distribution of all items in a brand, segment, category or any other aggregation of products.  Here is an example of one brand’s TDPs and how that relates to % ACV Distribution and Average Items Carried.  Let’s assume that Brand A is made up of 4 items.  The table below shows the % ACV Distribution of the overall brand and each of its 4 individual items.  Brand TDPs = 200.

TDP calc1

You can see that if a brand has more items and/or each of those items is in higher distribution, the TDPs go up.

Note that the brand % ACV is not used to determine TDPs.   TDPs are always calculated at the item level; including brand % ACV in your TDP calculation would be double counting.

For some products, TDPs and Average Items Carried will be redundant.  If your brand is in very high distribution (95% or more), then TDPs are essentially equivalent to Avg Items Carried – just move the decimal point over 2 places, as in this example:

TDP calc2

Bottom Line:  When you need one measure to summarize distribution, use TDP.  It takes into account how widely your products are available and how many items are available.  If your brand is in almost total distribution, you can use either measure.  In my experience, Average Items Carried is the better one to use since your audience is likely to find it more intuitive.  It’s easier to wrap your head around the number of items carried per store than a more abstract number like Total Distribution Points.

Have other questions about retail distribution measures?  Leave a comment below or or contact us and we’ll do our best to answer it!

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 twice a month. We will not share your email address with anyone.

 

PrintFriendly and PDF

Comments

  1. krishna Kumar Gupta says:

    Can you tell me how to calculate any specific Brand’s % ACV Distribution?

  2. Sally Martin says:

    Krishna,

    If you have distribution at the individual sku level but don’t have distribution for the total brand, there is no way to calculate % ACV Distribution for the brand. % ACV Distribution is a non-additive fact. But here are a couple of suggestions:

    1) Look at Total Points of Distribution (described by Robin in this article) at the brand level. This can be done by adding up the individual sku level distribution numbers. It’s not the same as % ACV distribution but it is something you can look at over time, across markets, and across brands.
    2) If any of the individual sku distribution numbers are close to 100%, you can safely assume that the brand distribution is close to 100%. In other words, brand distribution must be *at least* as high as its best sku.

    • Hi Sally,

      It looks like for calculating TDP, ACV Max would be used, as opposed to ACV Average, correct?

      Love the articles!

      • Sally Martin says:

        When using the % ACV measure, the choice of ACV Max or ACV Average depends on which you think is the most accurate measure of your distribution. And that judgement applies when you are using TDP as well (since TDP is a function of % ACV). So if you have a slow moving category with stable distribution and minimal out of stocks, Max is probably best. Or if you have a new product that is building distribution, Max also makes sense. If you have a quick moving category with product availability or out of stock issues, then Avg is likely a fairer measure of your true distribution.

        Any time you are thinking about any measure related to % ACV, ask yourself this question: “What is the fairest representation of my actual on shelf presence, given my particular business issue and product?”

  3. Marie-Therese says:

    Hi Robin/Sally

    Really great blog here, very insightful!

    I’m based in the UK and have never heard TDPs mentioned, however we do have something called cumulative distribution – is this the same thing?

    Best regards

    MT

    • Robin Simon says:

      I am not familiar with the term “Cumulative Distribution,” but it may be the same as TDPs. You can figure out of they are the same by doing the TDP calculation manually and comparing the result.

      An alternate definition for Cumulative Distribution could be related to the time dimension. Here’s an example: %ACV Distribution for a 52-week period is 45% but Cumulative Distribution for that same 52-week period is 72%. This would mean that the “regular” %ACV Distribution measure is the average of all the individual weeks but the Cumulative Distribution is the %ACV in which the product EVER sold over those 52 weeks. My example would happen for a new product – avg weekly distribution over the course of its first year could be 45% but over that first year it eventually sold in 72%.

      If you pull “regular” %ACV Distribution and Cumulative Distribution for a 52-week period, you should be able to tell which of these is the correct interpretation. I hope this helps!

  4. Gerri Spain says:

    Loved this post! Prior to reading this, I’ve been using Average Number of Items as the primary measure of breadth and depth of distribution. Now I find myself creating the aggregate for TDP for a more complete understanding of the true number of points of distribution. Much more useful for me to use the TDP results to truly understand velocity at the Brand, Segment or Category level.

  5. Hi Sally,

    Does the %ACV matter if it is Max or Avg when making these calculations? Or are you saying %ACV Distribution is a fact?

    Thanks,

    Bob

    • Robin Simon says:

      The ACV measure available depends on which tool you have from IRI or Nielsen. If you are using the web version from either supplier (IRI Market Advantage or Nielsen Answers), I’m pretty sure there is one key measure for ACV Distribution. That’s because the ACV-related measures are always calculated from the lowest of detail (UPC and weekly). Because you’re asking about Max or Avg ACV, I’m guessing you are on one of the Excel-based tools, IRI XLerate or Nielsen NITRO.

      Max ACV gives you the highest ACV for any of the items that make up the product total you are looking at and for the highest week in the period. Avg ACV gives the average of the items and weeks. So…if the product total you are looking at is pretty stable (ie – an entire category or segment or an existing brand), then both ACV measures will be very similar. If you’re looking at a new brand or segment then it’s probably best to use Max ACV, since using Avg would take the average of all the weeks of the introduction while distribution was still building.

      Hope this helps!

  6. Thanks for the great article! As I’m new to this industry, I have a doubt on aggregating ACVs. Say for example, I have 4 products/ SKUs under each product groups and I plan to analyse at the group level. I have data only on SKU level ACVs (not at the group level). Now I need to calculate the ACV at the group level. One method I thought of was to take the weighted average of the products’ ACVs based on the volume of sales that were made during the time period. And now when I come across this term TDP, I get a doubt whether to use this TDP or the weighted average of ACVs. Can you please explain me which is a better measure?

    • Sally Martin says:

      As is so often the case with data analysis, the choice of measures depends on what question you are trying to answer.

      TDP essentially tells you the average number of SKU’s carried per store in the market. It’s at least saying something about the group of products in aggregate.

      Your weighted measure tells you the individual SKU distribution achieved by your average SKU. It’s really not saying anything about the group of products in aggregate.

      Neither measure will tell you the %ACV that carries *something* from your line of products. That’s not possible with the data you have in hand. However, you do know that the % ACV for the whole group can’t be less than the % ACV from the item in your line with the best distribution. So if one of your products has close to 100% distribution, then you can assume that product group distribution is also close to 100%.

  7. Hello, thanks a lot for your articles, they are so useful and straight forward!
    Just one question on TDP: Which time frame would you advise to use?

    • Robin Simon says:

      The most common time periods to look at TDPs for are probably 4-week and 12-week. Sometimes you may want took at a longer timeframe and occasionally weekly TDPs can be helpful. If you are most concerned to see what your current distribution is, then I recommend looking at the most recent 4-week period. It essentially eliminates the possible understatement that can happen for slow-moving items that may not sell in every store every week. (You’ll see that the %ACV Distribution and TDPs usually increase the longer the period is. That’s because items sell in more places over a whole year than in an individual week.)

      If you want to see trends in distribution over time, you can look at weekly TDPs in a line graph. This will show if your distribution is generally growing, declining or flat. You may also see some seasonality in distribution this way, as retailers carry more items in certain categories at specific times of the year (for example, more kinds of candy in October or more kinds of sunscreen in the summer).

      Hope this helps!

  8. Hi and thank you for the great article!
    Is it possible to look at aggregate TDPs across custom aggregated IRI markets? Let’s say I’m looking at 4 markets aggregated and want to know a brand’s TDP in that new aggregate look; is this a possibility, or must TDPs be looked at only in individual markets?

    • Sally Martin says:

      Yes, TDP is an additive fact like dollars! That’s one of the things that makes it so useful. You can create a value for TDP from any combination of markets or periods or products. Just remember, though, that the more markets you put in your aggregate, the higher the TDP value will be. So you would need to take care when comparing TDP across custom aggregates. For example, if you have 4 markets in one aggregate and 5 markets in another, the TDP is likely to be higher in the second aggregate simply because it includes more markets.

  9. Hi,
    I am looking at a particular group of core upc’s and want to see what the total %ACV of just these 20 UPC’s would be for a particular retailer. I have %ACV, Sales Per Point of Distribution ($), and Dollar Sales. I want to add up the Dollar Sales/SPPD to get the %ACV of the 20 core upc’s.
    However, when I add up the SPPD for all of the UPC’s in the total brand the UPC total SPPD ($) are greater than the brand total SPPD$ that is delivered out of the Nielsen database. The total SPPD $ rolled up at upc level are about 47% higher in SPPD$ than the brand total that is delivered from the db.
    Can you help me understand?

    • Robin Simon says:

      Good question!

      You’re trying to determine what the %ACV Distribution is for a group of 20 UPCs, but those 20 UPCs are only a subset of the total brand. Unfortunately you cannot get exactly what you need, but here are some key points:

      – The %ACV Distribution for the 20 UPCs (i.e. % ACV selling any of the 20 items) will most likely be something somewhat higher than the %ACV of the highest UPC.
      – Think of SPPD $ for the brand total like a weighted average of SPPD $ all the UPCs that make up the brand, or the sppd $ for the average UPC in the brand.
      – You are correct that for a given UPC or subtotal on the DB that if you know 2 of the following facts you can back into the 3rd one: %ACV, Sales Per Point of Distribution ($), and Dollar Sales. The key for aggregates is that they must be on your DB, and not calculated.

      Based on the data you sent me, the %ACV Distribution for the 20 UPCs has to be something something between 89 (the highest distribution for any UPC) and 92 (the distribution for the whole brand). You could split the difference and call it 90.5 then use that as the denominator for the SPPD calculation.

      Hope that helps!

  10. Hi, Firstly Thanks for the great article. Very Useful.
    My question is “Is it right to say the highest TDP for a single SKU in a single Market, for 52 weeks will be 5200”?

    • Robin Simon says:

      Thanks for the question! The TDP for a single SKU in any geography is 100, for any period. At the SKU level, TDP is the same as %ACV Distribution so the highest possible value is 100. For any product aggregate bigger than a single SKU, there is no maximum, except for the total number of SKUs * 100 – that would mean that all SKUs are in 100% ACV distribution. The only way that the length of the period affects TDPs is that the value can be slightly higher for longer time periods. For example, the TDP (or distribution) for a SKU might be 87 in a 4-week period but 95 in a 52-week period. The less frequently that shoppers purchase something, the bigger the difference will be in the values for 4 weeks vs. 52 weeks.

      Hope that helps!

  11. What is Numeric Distribution Max? How to read these figures

    • Robin Simon says:

      Based on your question, it seems like you are looking at data from outside the US since we usually don’t see the term “Numeric Distribution.” I’m not sure if that means a % of stores or % of ACV or something else. Does the number range between 0 and 100? The “Max” designation means the measure is the maximum (as opposed to average) distribution achieved during the period. For example, max distribution for an item over a 52-week period may be 85 but the average distribution over those 52 weeks could be only 68%. Said another way, over the course of the entire 52-week period, the sold someplace in 85% but in an average week it sold in 68%. The longer the time period, the more different the max and avg distribution are likely to be.

      Hope that helps!

      • Ludivine says:

        it is the % of Stores (where the item or brand or manufacturer is present). so if the brand is present in 600 stores out of 1000 stores (for that channel) for example, then it has a 60% numerical distribution (irrelevant of how well the stores are performing).
        If it only says Numeric or Numerical Distribution: It is for all stores (ie the ones selling the products, or the ones simply carrying the product during the period)).
        but if it says “selling numeric Distribution”, it is for the stores who sold the product during the period (and not just the ones who carried it)
        the counterpart of % of ACV, would be Weighted distribution (WD), but most of the time WD is at Category level instead of all commodity volume.

        • Robin Simon says:

          Thanks for the clarification! (FYI, Readers – Ludivine worked for Nielsen servicing global clients so she knows this stuff!)

        • Hi, Firstly Thanks for the great article. Very Useful.
          LUDIVINE, would you please explain the WTD selling distrbution by the same way you did with “selling numeric Distribution”,

    • Numerical Distribution Maximum theoreticaly will be 100%, meaning for example out of 10 stores 10 have your product.

      Other term that I hear for ND is Trade Penetration Percent, so it might be therm used in US.

  12. Alejandra says:

    If I say that the Total Distributions Points increased by 10% versus prior year, what does this mean?

    • Robin Simon says:

      You can think of it as overall retail presence in the market has increased by 10%. At the item level, TDP is the same as %ACV so a 10% in TDPs means the item is in more stores and those additional stores represent 10% of the ACV. For larger product aggregates (like brands), a 10% increase in TDPs could come from:
      1. the brand getting into additional stores that it was not in at all and they represent 10% of the ACV
      2. more items getting into stores that the brand was already in
      3. some combination of these two (most common)

      Hope this helps!

  13. hi there,

    This blog is so helpful.

    Say that I have item ACV data for a market (ex US food over $2m) and I also have item ACV data at the account level.

    I know that it is accurate to sum the ACVs of each item in the market (US Food) and state this as the total distribution points in that market. Is it defensible to sum the item ACVs in each account (Albertsons, HEB etc), and then sum the TDPs of those accounts as a representation of distribution? Essentially, someone wants me to use the much larger number yielded from the second approach, rather than the former, which I feel is more accurate.

    Thanks!

    • Robin Simon says:

      While I like your creativity in coming up with another distribution measure, I do not recommend what you describe. If you sum the TDPs across retailers, then 400 TDPs in Kroger would get the same “credit” as 400 TDPs in a much smaller account, like HEB. You are correct that summing the %ACV of all items within a market is the most accurate and robust measure.

  14. Larry Belka says:

    If a chain has 100 Stores and 5 SKUS are authorized do I have 500 potential points of distribution?

    • Robin Simon says:

      Regardless of how many stores are in the chain, if you have 5 items authorized then the maximum potential TDPs would be 500 since the measure is based on ACV and not number of stores. TDPs of 500 means that 5 items are in 100% distribution. You may want to take a look at this post about Distribution which explains more about why “Authorized” and “In Distribution” are not the same thing as far as the database is concerned (http://www.cpgdatainsights.com/distribution/2nd-most-important-part1/). An item must actually scan to be considered in distribution. In your scenario, if all 100 stores are the same size (in terms of overall sales) and all 5 items scanned during the period, then, yes, you would have 500 TDPs. In real life, both of those things are rarely, if ever, true. Even within a chain there are usually stores of different sizes. Also, just because all items are authorized at the chain it doesn’t mean it is actually on the shelf in all the stores within the chain. I would guess that each of the 100 stores probably has 3 or 4 of the 5 items on shelf, although all 5 items are selling someplace within the chain.

  15. Thanks for such a useful effort
    My question is more on how to benchmark TDPs i.e. do you recommend a TDP range for Low, Medium and High performance?

    • Robin Simon says:

      Any benchmark is related to how many total items are available in your portfolio so there is no one answer for all brands. If you offer 10 items than the maximum TDPs would be 1,000 if every single store actually scanned every single item you offer (10 items x 100% ACV). If you offer 6 items, than that maximum would be 600 TDPs (6 items x 100% ACV).

      If you are comparing retailers to each other for the same brand, then you could look at TDPs for Total US and say any retailer that has the national TDPs + or – maybe 20% is Medium. Any retailer with lower than 80% of the national average would be Low and higher than 120% of the national average would be High.

      It is harder to fairly compare TDPs to each other across brands since they could have widely varying portfolios in terms of number of items.

Have a comment or question?

*