Most comprehensive measure of distribution, accounting for both the %ACV Distribution (reach) and Average # of Items Carried (depth, sometimes abbreviated as AIC). TDP is often available right on your database but it can be calculated for a brand (or category) by summing the %ACV Distribution of the individual items that make up the brand (or category). Example: Say Brands A and B both have 200 TDPs. If Brand A is in 50% ACV Distribution and has 4 AIC and Brand B is in 20% ACV Distribution but has 10 AIC, then Brand A is distributed more broadly but Brand B has deeper distribution in the fewer places where they are.
Learn more in this article:
Total Distribution Points: Master of All Distribution Measures
Hello,
I was wondering if it is possible to calculate item weekly average $ Sales Per MM ACV for a category. I would like to use the category avg. as a benchmark when I’m charting new item weekly $ Sales Per MM ACV. The issue is that I am comparing apples and oranges. The category weekly $ Sales Per MM ACV figure is much higher than any 1 item weekly $ Sales Per MM ACV figure. Is it possible to get an apples to apples comparison?
Thank You,
Joe
Hi Joe, In your database, the $ Sales per MM ACV for the category is for all items combined. In other words, its the velocity for the entire category, not the velocity for the average item in the category. As far as I know, there is no database measure for average item velocity but you could calculate that yourself to establish a benchmark. You might want to exclude items with very low distribution because sometimes their velocity estimates can be a little funky.
Would Units per store per week selling work for velocity? It doesn’t take into account the size of the stores that they are selling in, but can provide some guidance on velocity.
Yes, that velocity measure can be used. However, you cannot compare that measure across markets. So within an individual retailer, you can compare units per store per week for different items. But don’t compare compare the measure for retailer X to retailer Y. You can read more about different velocity measures here.
How do you interpret the TDP of 200?
For example, the %ACV of 50%, you would interpret as “Brand A is carried in 50% of the stores in the market”
Thanks for your question!
First off, on your % ACV example, the interpretation would more precisely be “Brand A is selling in 50% of the market ACV”. That may or may not translate into 50% of the stores.
TDP is hard to put into a comparable, interpretive sentence because you can calculate TDP across markets, products and over time. And, since it’s not bounded by 0% and 100%, there’s no intuitive feel for what the absolute number indicates.
Using your example, I would say: “In this market, the sum of % ACV across all Brand A’s UPC’s is 200”.
But I would never put that sentence in a presentation or report because I don’t use absolute values for TDP. Instead, I report how it changes over time (I use % change versus year ago), turn it into an average items number which is more intuitive, or index it so I don’t have to show the absolute value. Then if I think people aren’t familiar with the measure, I add a footnote with the definition of TDP: “TDP is the sum of % ACV across items over time”.
What does Base $ per TDP tell you?? and if the output shows a negative % change- what should I take from that?
Thank you!
Hi Julie,
Base $/TDP is one type of velocity measure (velocity is also known as sales rate). Velocity measures tell you how well you are selling, adjusting for distribution. If you aren’t familiar with velocity as a concept, I suggest you read some of the articles on our sight to address this measure.
In this case, your numerator is base dollars. Base dollars is an estimate of your sales without the incremental impacts of trade promotion. So a negative trend for this measure would suggest declines in your base business where your product is available – not something anyone likes to see in their data. What causes base dollar declines? Everyday price increases, competitive activity, reduced consumer promotion, and many other factors.
However, it’s important to remember that even though velocity measures adjust for distribution, that doesn’t mean they aren’t impacted by distribution changes. If you increase your distribution (a good thing) but your velocity in your new accounts is weaker than your velocity was in your established account, you overall velocity will go down. So when you investigate this trend, be sure you not only look into what might be impacting the numerator (base dollars) but also investigate trends in the denominator (TDP).
Great information. I understand TDP is additive. Trying to understand this better. I pulled TDP for a manufacturer in 27 different retailers. I also created a aggregate total for the 27 retailers. The TDP for the aggregate was not a sum of the aggregate (like dollars was summed). I didn’t expect that. It isn’t a straight average either. Wondering what it represents? TDP reach for the 27 are as follows but the aggregate value provided by Nielsen AOD data pull was 629.
668
338
517
437
684
328
613
1,090
343
353
365
595
1,048
682
497
364
393
849
569
741
314
779
674
583
636
281
1,184
Thanks for the question! TDP is additive across products but only within a market, not across markets. TDP for the aggregate of markets is like a weighted average across the markets, which is what the data you have looks like. Fortunately AOD does this calculation for you – in the older systems there was no way of knowing the true TDPs of an aggregate that was not already in the database. If you want to compare the 27 retailers to each other or to their aggregate you should do an index of each retailer’s TDPs vs. the aggregate. So, for example, the first retailer in your sample data would have an index of 106 (668/629 * 100) and the second retailer would have index of 54 (338/629 * 100). Saying that Retailer 1 has 6% more TDPs than the average and Retailer 2 has 46% (1 – 54) fewer TDPs than the average should make sense to most audiences.
Hope that helps!
I didnt get this post completely. What is AIC?
AIC is the abbreviation for “average items carried.” Other terms used that mean the same thing are: average items selling, average items, average number of items. Keep in mind that this measure (whatever you call it) is based on items that actually scan. So technically, “average items carried” is a bit misleading since a retailer can carry items on the shelf that do not sell in every store every week.
It is calculated for a brand by summing the %ACV Distribution of the individual items and dividing by the % ACV Distribution of the brand?
Now in the example above; 50*4 and 20*10=200. I am multiplying it here where is division?
Sorry but please elaborate.
Yes, AIC (average items carried) is calculated by summing the %ACV f the individual items and dividing by the distribution of the total brand. In the example, there are 2 different brands both with 200 TDPs. I’m just showing that one brand has much better %ACV (50 vs. 20) while the other has more items carried (10 vs. 4). The division comes in when calculating AIC since that is not always available as a fact on all databases. %ACV for items and brands is available so sometimes you have to calculate AIC.
Hope that helps!
Does the number of facings of any given SKU have anything to do with the calculation of Points of Distribution?
The number of facings is usually closely related to the measure “average items carried” (AIC). Keep in mind that you can’t really know facings unless there is a comprehensive audit where people actually go into stores and count the number of facings. A small number of companies do get this information but usually not more often than once a year – it’s pretty expensive to do a national audit. Average items carried is based on the items actually scanning in a store and is the closest surrogate for facings. See this post on the relationship between TDP, average items carried and %ACV.
Hope this helps!
Hi Robin,
I understand the TDP calculation. But lets say you have SKU level weekly data from Nielsen, for that TDP will be same as %ACV. But if you were to aggregate the 4 weeks of data at SKU level, TDP will be sum of %ACV for those 4 weeks. I know you can aggregate for Brand, I was not sure if you can aggregate it at SKU level across weeks.
Can you clarify this?
You are correct that you can aggregate across items to get brand TDPs or even add brands to get to category TDPs. I usually do not aggregate TDPs (same as %ACV at weekly item level) across weeks even for items. As far as I know, almost all databases have both weekly and 4-week periods available so you should be able to pull TDPs for the 4-week period instead of adding across the 4 weeks. Here’s an example: Let’s say an item has the following TDPs for 4 weeks – 54, 49, 52, 52. The sum of that would be 207, which is kind of difficult to interpret. If you pull TDPs for that same 4-week period you would probably get something in the range of 55-60. This is because during the 4-week period there is more opportunity for the item to scan than during any individual week. Sometimes the ACV or TDP for periods longer than one week is called “reach.” For longer periods there is sometimes a measure called “average weekly %ACV” which, in this example, would be 51.75 (essentially the average of the 4 weeks). Hope this helps!
What is TDP Reach telling me?
EX. TDP is 46.1, TDP % Chg YA is 7.8 and TDP Reach is 51.2……what is the 51.2 telling me about this item
If TDP and TDP Reach are different then you must be looking at a period that is longer than 1 week. (And since you are looking at an item, TDP is the same as %ACV.) You’ll see that all the distribution measures have 2 versions – with and without the word “Reach” in the name. The plain one (without Reach) is the average value across the weeks that are in the period and the one with Reach is almost always a higher number. So in your example, if you are looking at a 12-week period then the item sold (scanned) in 46.1% of the ACV on average every week but sold (scanned) in 51.2% of the ACV at some point during the 12 weeks. For items that re very fast-moving with short purchase cycles the 2 measures should be very similar. The longer the purchase cycle or if an item doesn’t sell every week, then the 2 measures will be more different, with the Reach measure always being higher.
Hope this helps!
Dear Robin, thanks for great article about TDP – the great measure for distribution. But, unfortunately I can’t use it because there is not real (even any) data for any channel’s %ACV in Baku (“strictly confidential information”?!). What can I correctly get about sizes of retail outlets that is floor square meter, which I think can indirectly point on (most probably) outlet’s dollar sales value. Could you please suggest can I use this measure (floor space) to calculate TDP to arrive at an approximate distribution in a product market. Thanks in advance.
Store ACV is used to weight distribution (rather than just looking at % of stores). So if you don’t have ACV, then your alternatives would be to 1) use another weighting measures or 2) just live with % of stores. I think square footage, if you have accurate data for that, would be a reasonable alternative measure for weighting stores. However, if you have doubts about the quality of that data or whether it properly reflects the economic horsepower of a store, then % of stores would be safer.
That being said, %ACV is a number calculated by the data vendors (Nielsen/IRI) using store level data. So unless you have store level data, you can’t really produce a comparable measure.
Hope this helps!
Thanks for suggestion. But if you mean in your 2) % of stores to use instead – the issue is that numeric distribution does not tell us about sales potential of an outlet, since supermarket with million turnover and grocery shop counts the same as sales points. Anyhow, if I take this measure how much an error should be considered in estimating TDPs of a category?
One more query – can you please give me advice what is the most important scopes of distribution analysis to conduct (that is what factors include such analyzes to have understanding about distribution for my company -FMCG/CPG). If there are formulas for calculations please point them. I will highly appreciate you help. Thanks in advance.
My point about % of stores is that it’s better than weighting with data that is inaccurate or incomplete. If you have perfect data on square footage, it would be better to weight by square footage. But if you have suspect data on square footage, I personally believe you would be better off with numeric distribution. At least you are not misleading yourself.
I can’t give you an estimate of the “error” introduced by using numeric vs. weighted distribution. I have no idea.
What to look at in distribution analysis? Here are the questions I would ask of the data: Where is my brand distributed and where is it not? What are the big opportunities for improving distribution (e.g. places where the category is strong or there is a good demographic for your product but you do not have good distribution)? How do I compare to competition? How deep is my distribution? Are stores carrying multiple varieties? What are the opportunities there? At specific retailers, what varieties are they carrying? Are they missing specific items they should carry (e.g. strong sellers elsewhere)? Do they carry items that seem sub optimal for them (weak sales rates or wrong variety for their demographic/geography)
We have a lot of article on the website regarding TDP and Distribution. I would read all those articles if you have not done so. On the home page, there is a category for “Distribution” which would list all those articles.
Thanks Sally
Hi. I understand this a very broad question, and depends on many factors, but is there a standard answer for correlation between total number of facings and overall average brand velocity? i.e. If a brand went from 2 to 3 facings in a chain of stores, you would expect the overall average velocity for that brand to increase, but is there a ballpark figure to say what percent increase that would be?
Thanks!
I don’t know of any estimates for this. I think there are so many factors that it would actually be dangerous to make assumptions. Sorry!
Hi,
Can $/TDP value be greater than the dollar sales value for 52WE period of a brand.
TDP is generally represented as a whole number. o assuming that is the case, then dollars/TDP couldn’t be higher than dollars. The only way I can imagine this happening is if TDP is mistakenly being shown as a decimal (because the underlying % ACV numbers are decimals rather than whole numbers).
Hi Sally,
I now understand TDP’s are not additive across markets. However, my leadership wants to see what SPINS reports as MULO and Natural Enhanced added together to be a total market. We are currently very strong within the natural channel and transitioning to conventional. Our number one competitor is the opposite and strong within conventional but does not have a foothold in natural.
How would you advise them to communicate a “total” business strength vs other brands to both internal and external stakeholders?
Thank you for your ongoing thought leadership in cpg analytics!
Thanks for the kind words, Amanda! You’re right that you can’t just add TDPs for MULO and Natural channel together to get the total distribution picture. Dollar sales is the most obvious choice to show how strong both brands are in total. You could compare the velocity within each channel fo the 2 brands. Hopefully your brand has higher velocity the the competitor in both channels, even though you have very limited distribution in Conventional. Check your email for a message from me about discussing this further.
– Robin
hi thanks for your great post!
now I am appointed to understand the TDP share and take actions to improve it.
but I confuse that whether the higher TDP% the higher market share%
first TDP% is a bench mark for distribution performance, but if the Market Visibility Index (%TDP share/ %Val share*100) is really high, (> 200%) is it still important for a brand to improve distribution. or need to improve SIH ?
another finding is that, in hyper which ACV is 100%, our brand market share grew but TDP share% declined, do we still need to take actions to improve AIC? is there marginal benefit?
We have a whole post on Fair Share of Distribution, which is what you call Market Visibility Index. I’m not sure what “SIH” is but you are right that it will be difficult to convince a retailer to increase your brand’s TDPs if the fair share is already very high, like over 200. To address your second question, it is not unusual to see 100% ACV distribution for the brand with market share of sales growing but share of TDPs declining. As long as every store sells at least one item in the the brand during the period being measured the brand %ACV will be 100%. If your market share is up but share of TDPs is down, that means that the items that lost distribution had lower velocity than the remaining items or that other brands gained TDPs while yours stayed flat. It only makes sense to increase AIC (avg items carried) if the the added items have velocities at least as high as the existing items. A better tactic might be to get more facings of the brand’s fastest-moving items. Hope that helps! If you have further questions, please contact us here.
Hello,
Using a L52 week dataset, can someone please explain what the difference is between TDP and Average Weekly TDP? I’m guessing Average Weekly TDP could be skewed lower than TDP if either of the below things are true:
1) A brand gained incremental distribution during the year. As an example, a brand might have had 300 TDPs on January 1st, but if it gained another 100 TDPs on December 30th, then the TDP would show as 400, but average weekly TDP would be 300? This assumes that my data set is L52W December 30th.
2) A brand had out-of-stocks on shelf throughout the year, which would causes average weekly TDP to be less than TDP.
Can someone please confirm #1 and #2 above are accurate? Thanks so much!
Take a look at this post which covers your question. I’m surprised your dataset/database says “TDP” with no modifier, like Avg Weekly or Max/Reach. In your 1), that describes TDP Max (or TDP Reach) and measures distribution ever during the 52 weeks, so the max at any point during those weeks even if it was only in the last week. Avg Weekly TDP is always less than or equal to Max/Reach TDP for any period longer than a single week and they can get further apart the longer the period. In your 2) you are correct that out-of-stocks to cause avg wkly to be lower but that also happens for slow-moving items with long purchase cycles where it’s possible for the item to be on-shelf but not sell in some stores during some weeks. Hope this helps!
Where is the measure “Average Weekly TDPs” within Nielsen?
I’m not sure what you mean when you say “where”? Do you mean you don’t see that metric in your database?
It’s possible that a reported TDP metric is an average, even though the fact is not named “average”. For example, I have a client with a Nielsen Byzzer database and, when I pull the fact Total TDP for a summed period of weeks, the average TDP is returned rather than the summed TDP (as I personally would expect).