Big Data? Big Deal! – PIM Is the Glue that joins Omni-Channel Marketing & eCommerce

Written by Brad Martell

For this brief I’m going to focus on PIM and will write at a later date about eCommerce Platform’s, CDN’s and DAM’s.

My original title for this post was going to be, “PIM, the ‘irrigation channel’ that sustains your multi-channel crops.” (Please forgive the Midwestern boys farming analogies):

I wanted to reflect how essential data, (like water) is to any organization but also how difficult it is to get it to flow steadily and efficiently to all the places you need it. Just like in agriculture, the engineering, planning, and organization that goes into laying out complex irrigation channels for crops is very similar when architecting a data syndication model that delivers your data easily and efficiently to all of the channels that allow your organization to thrive.  I use the word channels specifically to refer to things like print catalogs, outside sales teams, online web sites and last but certainly not least, channel partners who distribute your products.

Some Disclosure First

Try as I might to be objective and unbiased, I must reveal that I spent 20 years at an Industrial Manufacturer in Milwaukee, WI named Brady.  Therefore much of my opinions were formed during that time and I often write and comment from the perspective of a manufacturer.  That said I’ve also worked very closely with the largest national distributors down to the individual single location distribution outlets.  I’ve seen the perspective of channel and brand owner, and I think I understand very well the challenges big data represents to both of them.

Big data?…big deal!

I must confess the use of Big Data will bring down “hell fire” from my peers who hear me ridicule the rampant misuse and blatant overuse of the big data cliché.  Next to “the cloud” I’ve never heard any industry phrase uttered so frequently yet have as many different definitions or at least intentions behind it. For simplicity sake I will offer my definition for this post. Big data, relative to product information management means one or all of the following:

1. The sheer quantity of product data being generated by manufacturers/brand owners.
2. UGC or User Generated Content in the form of product reviews or social sharing platforms.
3. Video (enough said).
4. Product images, multiple image resolutions, and angles.
5. Units of measure associated with product quantities.

All of this and more add up to incredible amounts of data that need to be managed and “normalized” so it can be readily consumed by a variety of channels.

I typically talk about product data in the context of being structured or unstructured. The way I describe the two is, Structured data – that which could be put into an Excel spreadsheet; i.e. Price, Unit of Measure, Weight, Short Description, Long Description, etc.  Unstructured data – things like: images, video, reviews, technical data sheets, material safety data sheets, etc.

The Future of Big Data

With some guidance, you can craft a data platform that is right for your organization’s needs and gets the most return from your data capital.

Get the Guide

The point I’m trying to make is that for each product SKU a manufacturer creates, there is both structured and unstructured data. The other point I’m trying to make is that the sheer volume of this data is “mind-blowingly” difficult to manage…especially when it resides in several different data repositories.

Imagine the degree of difficulty for a single manufacturer who must manage all of these data elements for all of their sku’s.  At Brady I was responsible for a division that sold over 60,000 sku’s!  Now imagine a distributor (i.e. a Brady distributor), and maybe I only can manage to represent 20% of Brady’s sku’s on my website due to the sheer work effort….that’s still 12,000 sku’s! ….and oh yeah Brady is only 1 of 200 manufacturers whose sku’s I need to represent on my web site so multiply 12,000 x 200….are you understanding my definition of big data?!

Enter PIM or product information management

The overly simple way to define a PIM is that it is a single data repository for both structured and unstructured data.  It is often referred to as the “single source of truth” or the “system of record”.  Another way to put it is it allows a brand owner to take numerous disparate databases that house various product data elements and consolidate them  into one database.

An additional, and extremely powerful value proposition is the ability to create what I refer to as “flavors of data”.  What the heck is a flavor of data you ask?!  Read on:

Being a parent I used to love to be able to tell my kids, “you get what you get, and you don’t throw a fit”; while that may have worked on my 6 year old daughter it definitely doesn’t work when you are talking to a merchandising manager or product manager at Grainger, Home Depot or Lowes.  Each of them is going to “request” that the manufacturer provides their product data in a “flavor” specific to them.  We often refer to this as “product syndication.”  Any manufacturer who sells through big box retailers or the major national and regional chains will tell you that the standard product data feed that might placate the small retailer or the internal marketing team, will likely not satisfy the major outlets. “Fair enough,” if you have the influence to custom tailor your flavor of data to fit your specific needs from the manufacturer, why wouldn’t you? The alternative for these channels is take on the crushing task to normalize the data feeds from hundreds or potentially thousands of suppliers if you are a big box outlet.

For you manufacturers why wouldn’t you want to create that channel specific flavor of data for a big box outlet?  If you have any doubts about the incremental effort to this think about these stats:

• Those who implement PIM (providing specific data flavors) have their channels represent 40% more products online than those who don’t.
• Those who have more sku’s represented will see sales increase of 28% greater than those who don’t.
• Finally those who provide specific data flavors decrease the “time to market” in major outlets by 92%! This means if you provide your product data in the exact format your outlet needs, you will have those products published 92% faster than those who provide data in a format that needs to be manually manipulated/altered.
• All of the above means more sales, faster, and greater customer satisfaction to the manufacturer.

A little more clarity on what I mean by a “data flavor,” also known as data taxonomy. Each manufacturer has a specific flavor of data or data taxonomy for their products which they use internally. That specific taxonomy, that was created for the manufacturer likely won’t be able to be readily accepted by their channel partners. By way of further explanation I’ll use some specific, but common product attributes that can vary, but systematically can mean the same thing. Color is a universal example. A manufacturer may refer to the color “black” as the abbreviated “BLK” while the channel partner may call black BK, BLACK or BL.  Every business tends to develop their own internal coding mechanisms to define the same attribute, but they can be defined in very different ways. This can be due to limitations in field characters allowed in a business system like an ERP.  For example, at my former company we were restricted to 35 characters in our SAP system for a “short” product description. These system limitations required us to get creative with abbreviating words, dimensions, weights, and every other conceivable attribute for a product.  Imagine when we’d send our product data feed to our distribution partners. Many times they wouldn’t be able to decode our product taxonomy when they’d get something like, (1.0×2.5 WML, B292, UOM250, $37.5)  which was a 1 inch by 2.5 inch wire marker label, made of Brady 292 self-laminating vinyl, in a roll of 250 labels, for $37.50. What?! That wasn’t obvious to you?

Now imagine needing to not only make it apparent to all channel partners what a product description is, but also needing it to conform to their own system limitations.  This is the distributor’s greatest burden. They have to normalize product data from all of their suppliers to conform to their system requirements. PIM software sits as the system of record for the manufacturer or distributor and ensures data taxonomies are adhered to and cannot be pushed to a web site or a catalog without first getting scrubbed and being made to conform to the data taxonomy rules. PIM also allows you to create versions (flavors) of the same data set that conform to the taxonomies of a web site, a print catalog or a key distribution partner like a big box retailer. The PIM has you maintain product data in one place but simultaneously manage all the various flavors of data when changes get made to the primary flavor.

One of the real value drivers of PIM is the amount of time it takes to maintain all of the various flavors of data without a PIM versus with a PIM. One of our clients told us that prior to their PIM investment they would spend $29,000 (“calculated in man weeks”) getting their product data to push to their largest MRO channel. This MRO channel published their product book every other year and their success as a business was largely determined by how accurate and how many SKU’s they could get into the “big book”.  After the PIM investment and investing in the time to create a specific flavor of data for this key channel the amount of effort went from $29,000 of effort to literally minutes. They also increased their SKU count by 34% in the big book and were asked to bid another product segment which they previously had not sold through in this channel. What they were also able to avoid were the fines associated with providing bad product data (i.e. the wrong flavor) to the big boxes. That’s right…fines. One national big box has taken to levying fines. For every cell that is incorrect the manufacturer gets fined $100/day/cell.  Imagine if each sku had 100 cells of attributed data. Now multiply the number of sku’s and if one cell is wrong for every 10 sku’s, the fines could certainly mount up.

Those fines would look small to the manufacturer who got replaced by a competitor because their data was improperly formatted and they couldn’t comply with the big box’s request for their flavor of data. This is happening today…this isn’t theory.

I don’t want to end on this dire note.  Manufacturers or brand owners should focus on the opportunity, not the penalties for not complying. My advice to the brand owners is:

• Invest in PIM.
• Enable your channel partners…
• Sales will grow.
• Time and effort will be saved internally.
• New products will get promoted more quickly.
• Show a commitment to those partners who have desire and ability.
• Leverage PIM to streamline your catalog production as well.
• Make sure the PIM serves your internal channels first like web, catalog, and sales.

Another bit of advice is find out how many different data depositories house the various data elements for your product. Calculate how much time is spent managing and maintaining data, and then ask your key channels if they’d represent more of your products if you could create a specific data flavor for them. When you get the answers to these questions you’ll have the framework to cost-justify this investment. Which by the way can be as little as $100,000 or over $1,000,000 depending on sku counts, numbers of channels, and if print catalogs are in the mix.

Thanks for reading and please give me your comments around managing product data, and provide your own stories about the value you’ve found in implementing a product information management tool.

As I said my next post will be about CDN’s, DAM’s and / or commerce so please, keep tuning in.

All the best

Brad