With Susan Walsh, The Classification Guru
00:00:00 - Introductions, book-writing, and passion for data
00:13:00 - Data as an investment rather than a cost
00:21:00 - Common visibility issues in data classification
00:27:50 - Hidden IR35 risks and cost-saving opportunities in the tailspend
00:33:20 - Standardising practices and data in procurement
00:42:50 - The role of the data detective
00:48:00 - Data best practice for procurement
00:52:00 - The business case for good data practices
00:55:00 - Predictions for the near-future
Jonny Dunning [0:00]: Excellent, we’re now recording. So, I’d like to extend a very warm welcome to Susan Walsh also known as The Classification Guru. Thank you very much for joining me.
Susan Walsh [0:11]: Thanks for having me.
Jonny Dunning [0:13]: So listen, you’ve been extremely busy of late. I know you’re always very busy.
Susan Walsh [0:11]:[inaudible] [cross talk].Yes.
Jonny Dunning [0:19]: So, first of all, what’s been keeping you so busy just recently?
Susan Walsh [0:25]: Oh, well, the biggest thing that’s been dominating my life for the past month or so has been my book. I had a deadline of the 2nd of April and January. I maybe had about 5,000 of the 50,000 words I needed. February, I decided to move hives, so didn’t get much done at the beginning of February. By the end of February, I maybe had 10,000 words and then March was just a mad dash to get those other 40,000 words done and most of those were done in the last two weeks as well.
Jonny Dunning [1:02]: Is this where the photo of you on LinkedIn with your both wrists strapped on –?
Susan Walsh [1:09]: I’ve still got one strapped up now, the other one’s not as bad. Yes, I’ve pounded them hard like literally 12/14-hour days, seven days a week for a month. They are feeling the pain now. But you know what? I can’t believe I wrote a book. Me. I wrote a book. [1:29 inaudible] a book. Yes.
Jonny Dunning [1:31]: Well, I think it’s a fantastic achievement, hats off to you, that’s absolutely brilliant.
Susan Walsh [1:36]: Thanks. Yes.
Jonny Dunning [1:37]: Just to get started with you on that basis, do you want to give us a bit of a kind of walkthrough of your career? Because I know it doesn’t necessarily start off with this area, just in terms of how you got to where you got to and what’s your journey’s been along the way?
Susan Walsh [1:53]: Yes. So, I went to uni, did a degree in commerce, wasn’t really quite sure what I wanted to do when I graduated, I ended up becoming a paint merchandiser. I’d go around being [2:07 inaudible]at DIY stores, merchandising paint, making sure that there were nice and tidy on the shelf, do a planogram; did that for about 18 months two years and moved down from Scotland to Guildford with that job then decided to try something else.
It can only be described as a horrific experience in recruitment before turning to sales. And then I went into telesales for Phillips, which got me into telesales at Colgate, which got me eventually into an account manager role with Dental Wholesalers. So, I was selling toothpaste, mouthwash, etcetera to them. I really did love that job. I was really good at that.
But after about two years kind of felt like I got as far as I went, had a couple of offers from other departments within the company, took the one that was more salesy, so I went into a national account exec role, ended up being more of a marketeer, shopper marketing, which looking back now has been great for my classification experience. But at the time I was like, “This is not what I signed up for” so then I went to try account management at the national account level, but with SC Johnson, so I did that for a little while.
And wrist problems are not a new thing to me, I ended up having to have a wrist operation, and so I couldn’t end up doing the job, it was desk-based, I couldn’t drive to the office, so I left and that’s when I decided to open a shop. So, that was a women clothes shop and beautiful shop, lovely clothes, well designed, but because it was in Guildford and it wasn’t [4:07 inaudible], people weren’t interested.
After about seven months I had to shut that up, had completely run out of money, wrapped up a ton of debt, had to go bankrupt but couldn’t afford to go bankrupt because it cost 650 pounds. So, I had to get a job so I could save up to go bankrupt. I found an ad online and that’s where my whole data procurement journey began. I started classifying spend data for a spend analytics company, and just at that point, all those different random jobs that ever had just came into one perfect formed ball of experience that I could use and I just found that I really enjoyed doing it.
I was good at it, I was efficient, and yes, I had a great five years there, grew as the company grew, I ended up managing a team there, running the projects, making sure everything was classified on time, showing the standards, the quality; and then it kind of got to the point where I was like, I can’t really go any further. I’m not learning anymore. I really love what I’m doing, but I don’t know where else I can get a job doing it, I didn’t train in data. I’m not an analyst. I don’t code.
I’ve never worked in procurement; I wouldn’t know how to buy anything if my life depended on it and I didn’t even know what the job I was doing would be called in another company. So, at that point, I made the decision to start The Classification Guru, so that was nearly four years ago, that’s where it started. And rather than doing the nice, sexy, glamorous, but the analytics and the cost-savings and analysis, I’ve always focused on the raw data and cleaning, and prepping, and classifying up.
And from the minute I started exhibiting and going to events, people said it was a great idea, “Oh I wish I’d known you six months ago” but because it was a service that Wasnt really offered, nobody knew to look for me. So, I discovered pretty quickly I had to make myself be heard because trying to reach out to other people was going to be quite hard to find the right people.
Jonny Dunning [6:36]: Did you feel like there was a requirement for a bit of market education?
Susan Walsh [6:40]: Oh, absolutely. It was the people who knew how valuable the service was but then, I also heard, oh, I’d never get that signed off, that’s the kind of attitude. So, I started to try and show the value of data through lots of videos and posts on LinkedIn and I’ve been doing that pretty consistently for the last two years specifically.
Which has really helped not only get people engaged and involved in the discussion, but talk, share their own problems and which in turn, shows that there is a wider problem and hopefully, some of the decision-makers are watching, lurking in the background, and are taking it in. And I would say that is definitely the case because all my clients now, they’ve all approached me because they’ve seen something of mine on LinkedIn at some point.
And it may not be that they contact me immediately, it could be a year later, it could be six months later, but seeing my post consistently trigger something that they know that I’m the person to come to and they’ve got that kind of problem. And I guess I haven’t really told you what I do, so that would probably be quite helpful. I specifically specialize in spend it classification and what I’ve found is that actually a lot of clients now would like built customized taxonomies for their business.
The off-the-shelf days, the UNSPSC with all its words and not the most user-friendly terminology and things like that, those days are kind of diminishing and businesses are looking at what’s right for their business rather than a kind of generic off-the-shelf package. So, that’s what I spend a lot of my time doing.
Jonny Dunning [8:44]: I find it fascinating because you’re doing something so specialist and you kind of fell into it –
Susan Walsh [8:51]: Oh, completely.
Jonny Dunning [8:52]: But there must have been elements of your character and your natural abilities that led towards these kinds of very organized approach. There must have been things that you would -- you mentioned kind of like some of your early roles, it sounded like there was a bit of organization going on with those and it was a combination of all those experiences, but you’ve almost created like a very specific category in some ways in the sense of –
Susan Walsh [9:19]: Yes, it did. This kind of -- not the job, but I guess the industry, the path that I’ve created, I have forged, that path wasn’t there before because the cleansing that was always kind of an afterthought. It was thrown onto the side of another service or another product you buy –buy this software. And I moved [9:46]we’ll tidy your data for you, but it’s like more of I kind of dust around the edges rather than under the table service, we’ll tidy up the main spend, but we might not necessarily look under the hood too much and see what’s going on there.
Jonny Dunning [10:05]: Sorry, I was just saying when you consider it, it’s crazy because if the underlying data isn’t clean, isn’t organized, then the results you are going to get are going to be very, very fundamental.
Susan Walsh [10:16]: And I think that there’s a bit of miscommunication on how much the third-party providers offers and does and what the expectation of the client is. I think they expect the whole shebang whereas the suppliers will do the minimum that we can get away with, but not always, but sometimes.
Jonny Dunning [10:39]: It’s not their core business, is it?
Susan Walsh [10:41]: No, but they just need to get it done to get their job done. As it turns out, it’s a big thing and I’m not creating some amazing new technology or building a rocket to go to space, I’m just classifying data. But it’s the bit that’s the most important but is the least paid attention to and I write about this in the book as well. You go and look at universities, you get all these students who are massively qualified including working with AI machine learning, but I bet you none of them have been trained on how to actually cleanse the data before they work with it.
And most of them will have worked with Kenneth Clinical Sterile Dataset, they’ve never really had their hands on a real-life dataset where nothing matches up, nothing’s named correctly, it’s all formatted wrong. They’re not getting trained on that, so I’m trying to highlight that this is the bit that’s the most important. It’s kind of viewed as the most menial task right now, but actually, it’s the most important because nothing else will work if your data is wrong.
Jonny Dunning [11:59]: It’s the groundwork, isn’t it? It’s like you’ve got to decorate a room. It’s all the prep that takes 80% of the time. I love the phrase you use, you talk about dirty data, and so your book if I believe I’m correct in saying it’s called Between the Spreadsheets, Classifying and Fixing in Dirty Data. Is that correct? When is the book coming out by the way?
Susan Walsh [12:23]: Don’t know yet, sometime in the next six months, so watch out for that. I will be shouting about that loud and proudly that’s for sure.
Jonny Dunning [12:30]: Thanks. I’ll be keeping an eye out for that. But this concept of dirty data, I really like it because I’ve seen this firsthand in businesses that I’ve worked in where even just at the level of the taxonomic classification, it’s all over the place and that’s a major problem. Put, for example, Salesforce into an organization and you look at the migration and the stuff being changed over many years, different systems have been used, or there’s the classic of an in-house system which just gets completely organised organically.
Susan Walsh [13:00]:[inaudible].Just matched data, and not bother, did you duplicate it first?
Jonny Dunning [13:05]: All of this stuff and you’re basically in a rare position where it’s kind of like if you work at a really nice hotel, but you get to see behind the scenes. When I was at university, I worked at a five-star hotel, it might be my holidays and you’d see behind the scenes as to what goes on. So, you have this beautiful façade and then things are always a little bit more gritty behind the scenes.
Susan Walsh [13:26]: Yes, effing and jeffing in the kitchen.
Jonny Dunning [13:28]: Exactly. So, where that applies to data it is a bit of a dirty secret, no one really wants to admit that they’re disorganised and no one wants to admit that maybe the data that they’re going to base their analysis on isn’t that great, but clearly, it’s a very real thing. I mean from our angle, obviously focusing on services procurement is something where we see it is just a massive lack of data, but certainly, in other areas you deal with, there must be huge amounts of data, but it’s just all over the place.
Susan Walsh [14:03]: Also, the first thing is it’s also not valued as well. So, I’m always constantly saying that your data is an investment, not a cost. So, there’s no incentive for a lot of people at the moment to look after their data or maintain it because nobody at the top bothers, then why should we?
Jonny Dunning [14:30]: Why do you think that is?
Susan Walsh [14:34]: Again, it’s the perception that it’s a menial task, that’s something that I shouldn’t could do, we could get in to do, and it’s not like that at all, you have to know the data, you have to be able to read the data and look for patterns and especially if it’s procurement data you need the context and knowledge. I know from training people that that can take three to six months because if you’re in a company, then it would take less time because you’re working on the same data each time.
But in a business like mine where each client is a different industry, then it takes longer to learn because every time you get a dataset, you’re learning a new industry from scratch. And yes, it’s just they don’t understand the value that it can bring beyond just it being tidy. It’s the decisions that are made, it’s the time spent by most of the organization fixing what’s not right, you present to the board and your charts are wrong. Well, that’s losing credibility within the organization. Nobody’s going to trust your work or your numbers and it’s so much wider than just all the data needs fixed.
Jonny Dunning [16:01]: And yes, it’s strange. Obviously, two things that I was going to mention there. Firstly, when you’re dealing with different companies and different types of data, clearly, you can apply some common principles across that which is this is where you get the coat on in terms of the –
Susan Walsh [16:19]: Ah.
Jonny Dunning [16:20]:Ah. See. Well, I like the fact you went from a clothing shop to a data business where your product is the coat that keeps the data nice and cozy.
Susan Walsh [16:32]: And that all started because of Daniel Barnes, I was going on his podcast about a year ago actually and he’s like, “Oh, you know, you really need a good acronym” and I was like, “Okay, well, data really needs to be consistent” so I think I had my C and then I had maybe my T and then the rest came. I think accurate was next and then I just needed an O and then I was like, “Okay, organised”. Yes.
Jonny Dunning [17:05]: I love an acronym.
Susan Walsh [17:07]:Yes, but then I was like, you can have so much fun with it. And actually, a lot of what I’m trying to do is not speak to the data people because that’s preaching to the converted. It’s trying to make data more accessible, more fun, less intimidating to the people who work with it, but are not necessarily experts because getting them to understand how important it is to input that data right the first time is the hardest job.
Jonny Dunning [17:35]: But you sort of think that people would look at Google, Facebook, and these massive companies that control huge amounts of data that have enormous power because of the data that they control rightly or wrongly or scarily or not. You think that that would just make people appreciate it more. Data is value. Is it the most important commodity these days?
Susan Walsh [17:59]: Well, I think within the data there’s probably a hierarchy there because of course, personal data is valuable, you can get insights into people and consumers and their behavior. But financial data is supposed to – not everybody sees the importance of that. And also, you’ve got GDPR are in place or Regional Data Protection laws for personal data, but there’s nothing in place for other types of data, so it almost puts personal data on a different level whereas all data is important.
Jonny Dunning [18:44]: I totally agree and from a business point-of-view surely the plight of procurement within the wider organization and having a really powerful and active role strategically within the organization is going to be to a large extent built on data. Procurement have a lot of data effectively going through their control, whether it’s on the services side, the same a lot of that’s kind of not really captured in many cases these days, but its huge potential for power, and data, and the ability to provide insight to make decisions.
But for example, in a manufacturing organization, all the information around suppliers, and goods, materials, and things like that is hugely powerful information, but if it’s unreliable or if it’s – I would imagine in a lot of cases just unusable in terms of being able to want to apply a really good data inside its packages and AI to the data. If it’s all over the place and the classification’s all different, you can’t really get started.
Susan Walsh [19:47]: Yes. And sneak peek from the book, something that I talked about in that is with product information. If you put the wrong dimensions of a product in and then the AI works out how many you can fit on a pallet and that’s wrong and then it gets to loading the truck and suddenly you can get a fifth of the pallets that the AI thinks you can get, your deliveries are messed up all over the place.
And actually, I don’t think it quite made it into the book, but I put out a request for some anonymous data horror stories and I’d already written this chapter about having the wrong dimensions and someone had written a story saying that they had like, it was one by one by one so they thought they could get like 3,000 TVs on a pallet, something like that. And we laugh, but this happens all the time like this is not uncommon.
Jonny Dunning [20:46]: Yes, it’s just kind of the mind boggles a bit really. But one of the things I thought was interesting to talk about was just the visibility because that’s a crucial part of it. And you’re one of the few people that gets to see behind the scenes within organizations where they’re basically just opening that kind of messy bedroom door and going, yes, here is everything and it’s full disclosure. What do you see in terms of common issues around visibility and the problems associated with that?
Susan Walsh [21:26]: Well, the first thing is I don’t judge, so I might take the mick out of them a little bit for some of the data and some of the typos and misspellings, but I think the first thing is to make them feel comfortable that this is actually normal and this is actually what I do for all my clients and there are loads of people out there with those issues. Things that I see a lot are missing information, lots of typos and misspellings, not usable invoice appeal descriptions. I’ve seen a lot of supplier names misspelled, which is –
Jonny Dunning [22:12]: All supplier entries you get all that sort of thing.
Susan Walsh [22:15]: I think at one point I did even find like IMB instead of IBM, like that level of carelessness and not paying attention to detail. There’s lots of it and then some have classified data, some don’t, but then they haven’t really updated it properly or nobody’s looked after it so it’s fallen into disrepair. And data is a bit like a classic car really, it may be old but it’s only valuable if you maintain it. If you let it rust away, it’s absolutely useless.
Jonny Dunning [23:01]: Yes, that’s a great analogy. So, obviously, you can’t classify data if you’re not tracking it, you need to get visibility, you need to have things in the right format, matching up, and actually available. If you look at that as how that within procurement and in terms of how that applies to procured services, what have you seen happening specifically around that area?
Because from our point-of-view, around procured services, we often find that there just is a very, very limited amount of data in the sense that you might have a statement of work contract that’s created that may or may not have milestones at what’s can be delivered, which may or may not have changed in the delivery of the project. But in most cases with services, it’s far harder to track and companies are much less likely to track it because existing technologies may be more geared towards goods and materials.
Susan Walsh [23:54]: And a lot of the services can be just people’s names as company names. So, it’s not necessarily ABC Consultancy, it could be Bob Smith. Those types of names can often be misclassified as employee expenses and things like that, so it might not even be necessarily categorized in the right area. I’ve also worked with clients where there is no description and I’ve had to just work with the supplier name in multiple languages, which was an absolute nightmare.
At that point, you can only do your best guess in some situations, okay the company I’m classifying for does this, this company could do this or it could do that. What’s the most likely service the clients are going to be buying? And then you have to take that judgment call. So, yes, it can be pretty messy, and then you’ve got your consultancy services, your big [25:06 inaudible], just seeing things like PWC p.w.cpricewaterhousecoopers price space bar house coopers and all the variations that go with that.
They’re not even getting a true picture of what they’re spending with one supplier let alone the whole business. So, that’s common and possibly one other common thing I wanted to just add in as well was that procurement have to kind of look at their data separate to the rest of the business because if they try to merge it with the finance system – finance isn’t happy because the categorizations aren’t right for them, master data don’t like it because you could have multiple classifications against one supplier and they only want one classification.
So, the work that procurement do in my experience tends to be done as a kind of offshoot of what’s going on in the wider company in terms of data because they can’t get what they need from the wider business.
Jonny Dunning [26:09]: So, how does that then all loop up if procurement are having to segregate their data to get what they need out of their information, is it then procurement’s responsibility to then feed that data back into master data and finance or is it the reverse of that process?
Susan Walsh [26:26]: It stays isolated or siloed within procurement for what they need it for. I did some work for a client and classified their data for the first time, we looked at one supplier that was a car leasing company, but it had four GL codes against it. And you can’t solve those problems. I mean highlighted that maybe some more work needed to be done GL classification in finance, but you can’t really --
Well, you could have extra columns, it really depends on how the system runs. But the way I tend to see it is that they just share the output with the rest of the business and that’s normally in some kind of chart or visualization, but the actual heavy lifting and all the analytics behind it is done within the procurement department.
Jonny Dunning [27:25]: Yes. It’s really interesting because again, that shifts a lot of responsibility under procurement or places a lot of responsibility on the procurement, but it’s also very empowering [27:37 inaudible] because they’ve got such a wealth of information. It’s interesting what you said about when you’re looking at suppliers and maybe it’s the service but it’s under the supplier’s actual name.
That’s really interesting and obviously, when you get into tail spend the risk side of it is massively increased and that’s a critical area. You’ve got lots of small suppliers, but then you’re into stuff around like employment misclassification as well with IR35 having gone live yesterday in the private sector. Is that something that you see come up?
Susan Walsh [28:14]: Yes.A couple of clients that I have used a lot of freelancers, but that’s changing and even just with that speaking to friends who work in say banking, they’ve had to get rid of all their contractors and now they’re expected to do the work of the contractors and they’re like, “We can’t do this” so I can see a lot of issues coming up in the next 12 months where they’ve lost that skill and knowledge and it could be within procurement, it could be within IT, it could be within finance, but I can see a lot of things going wrong potentially.
Jonny Dunning [28:59]: Yes. I think that’s an area where there’s a lot of dirty data, procurement services tail spend and there’s a significant risk specifically in relation to IR35. And I think a lot of people are taking the IR35 risk head-on as they should be and looking at their contractor populations, but what they’re not necessarily doing the in first instance, is also looking at their services procurement tail spend to understand the hidden headcount where it is within that whether it’s [29:28 inaudible] …
Susan Walsh [29:28]: [inaudible] search.But you know what? I’ve seen yoga teachers in there, I’ve seen math tutors and there’s such a range not yet, not necessarily the kind of services that you would think as in a consultant in lots of different types of services that are hidden in there. And again, it’s like I was saying at the very start when you buy the software or the service and that company says I will clean it for you, they’re looking at your top 80/90% of spend there. You dig into that tail spend and you will find a whole load of things that aren’t quite right.
Jonny Dunning [30:12]: And do you think there’s an element of additional risk around, obviously, this rogue spend, but fraud is something that does come up occasionally where stuff –
Susan Walsh [30:25]: Yes, sorry.
Jonny Dunning [30:27]: I was just going to say, is that more in terms of the way something’s classified or is it just kind of like unearthing stuff that shouldn’t really be happening in the way it is?
Susan Walsh [30:37]: There’s a couple of sites. So, there’s the supplier side of things where you might be buying things through that supplier that you shouldn’t be or maybe I know like from TV programs I’ve seen how there was fraud in the NHS where they got, the PO page twice because one time it was a zero and then the next time it was an Othat hadn’t been picked up.
So, there’s that kind of thing going on and then you’ve got employee expenses and there’s a lot going through employee expenses and not just fraudulent stuff, but I see a lot of software going through there. So, Adobe, MailChimp, SurveyMonkey, Microsoft, and that’s all influenceable spend, that’s not being accounted for because it’s going on someone’s credit card. So, there are all those kinds of things that you can start to pick up only to be classified properly.
Jonny Dunning [31:36]: Yes, I guess it’s not necessarily always people trying to do things the wrong way, it’s just a question of people just trying to get stuff done.
Susan Walsh [31:43]: No, sometimes it is a case of like we just need this software, put it on the company credit card, but I have seen instances, where all that spend and that could be hundreds and thousands of rules, is literally classified as employee expenses and that’s it. And that could be millions or billions of pounds/dollars whatever. But actually, within that you’ve got travel, you’ve got software, you probably got professional services, you’ve got office supplies, you’ve got catering.
People are buying the milks of the fridge off of – and it’s like there’s so much hidden information that tells you not just about what’s going on with spend, but what’s going on with your business, you can see where they have like drinks, Fridays, or they have like social – you can tell the companies that look after their staff because you can see it in their expenses.
Jonny Dunning [32:38]: Yes, like you say it doesn’t need to be a dirty secret because most people have got these problems, most companies have got these problems, and actually, people should be proud that they’re tackling it and taking it head-on because ultimately, if you can see that data, it’s just going to provide you with so much value in terms of just looking retrospectively, making predictive analysis, and just understanding your business in more detail.
Susan Walsh [33:04]: It’s that thing, everyone thinks they’re the only one, but actually everyone’s the same.
Jonny Dunning [33:10]: Yes. I can definitely imagine, as I said I’ve seen plenty of examples of this. So, another thing that I thought would be interesting to talk about was around the lack of standardization, and so that’s something that we see a lot in services procurement in terms of how people do their contracts, how tightly they monitor things like milestones in the delivery of services, but you must see that on a very much broader basis in the sense... We see it when people say, can you integrate with our finance system for invoices, for example? Yes, no problem. And we see people using the same finance systems that are effectively using them in very very different ways and obviously, deployments of that sort of thing can happen differently, but what sort of stuff do you see around that whole kind of lack of standardization side of it?
Susan Walsh [34:00]: A lot recently, and I’m talking across the globe not just specifically the UK. A lot of clients don’t have a statement of work template or an NDA and we’ve had to use minds and modify that and I’m really surprised at the number of organizations out there that don’t have those kinds of contracts in place because when I got my solicitor to draw them up last summer, I was like,” Oh no, I don’t need a…, I just wanted like me to my contractors’ NDA and me to my contractors’ statement of work contract.” Because he offered to do me like me to the client and I said, “No, I don’t need that. I always just sign there and it’s fine.” And of course, as soon as I said that, the flood gates opened and suddenly everyone’s asking if I’ve got a statement of work, so I was like, “Okay so, I always assume I’m the least smart person in the room with the least amount of contracts, and knowledge, and things”, but actually, it’s not always the case and it’s not always small businesses either.
Jonny Dunning [35:13]: No, absolutely. We definitely see that. I think it’s an important element though because where a supplier is working on something very, very specific and they’ve got a particular skill set, they can almost write the requirement of what needs to be done better than the end-client in a lot of ways. So, I think, certainly, within the way that we operate with customers, the supplier is very much part of the requirement building and agreement process then sucks that information into the template statement of work, which would generally be the client’s template statement of work their lawyers have approved because then–
Susan Walsh [35:49]: Well, that’s the thing, that’s what I assumed was that it would have to be absolutely grilled by their legal department so they would have that already.
Jonny Dunning [35:56]: I think they usually do, but it’s just whether people are using it or not and whether it’s enforced and particularly if you get into tail spenders below the procurement threshold where it’s getting that major scrutiny, then there’s not so much control around it, which is obviously from our point-of-view something that’s controlled by using the system and having also been going through the system.
Because naturally, you’re using templates – makes it easy for everybody, but it also allows the supplier and the client to collaborate and create the requirement because otherwise, you end up with either no contracts or you end up with a statement of work contract that just goes in a dusty shared file somewhere. No one ever looks at it again, and it’s never really measured against, so by the time the project’s delivered, it’s changed, it recognizably from that contract.
Susan Walsh [36:40]: Yes. Sounds familiar, yes. I’ve done my part as well, but most of the time I don’t even have like a proper sign-off process, it’s just that, you’re happy with that? Yes, okay, that’s good. Just give me a call if you’ve got any problems, and that’s how it ends. There’s no official ceremonial parting of ways or signing off of this a lot. It’s just that, all right, see you later.
Jonny Dunning [37:06]: And with that kind of completion of your work, you obviously mentioned that you weren’t necessarily primarily dealing with data people to start with because you’re preaching to converted there.
Susan Walsh [37:19]: Never did this to people, it’s always procurement that come to me.
Jonny Dunning [37:23]: Why is that? Is that because the procurement people have got the ultimate responsibility for it? Or is it because the data people aren’t given them enough of a voice within the organization to…?
Susan Walsh [37:35]: I don’t think the data people understand what procurement needs. I think that procurement has probably tried to get what they need internally before coming to me or someone else but has struggled, or they’ve been told, oh, yes, we can roll that out in about 18 months as a program blah blah blah, and they’re like, no, we need this now. So, they’re procurement, they can sign off their own spend so they can take action.
Jonny Dunning [38:07]: So, what is it that they need that they’re not getting?
Susan Walsh [38:12]: I don’t think that the rest of the business understands what procurement does, how these data or the skills involved in it. And again, it’s down to that, data is not really valued and oh, anyone can do it. But no, especially within procurement, the level of skill and knowledge that some of these buyers have it’s crazy and the negotiation skills and things as well, there are so many different aspects to procurement.
Jonny Dunning [38:52]: Possibly been undervalued.
Susan Walsh [38:54]: Yes, but maybe they need to step up a little bit too and promote within what they’re doing as well. I think it’s an all-around thing, it’s not one-sided.
Jonny Dunning [39:06]: It must be very frustrating for procurement people who are caught in that situation where they’ve got expertise and all of the insights to apply to a dataset. But if they can’t get the information and it’s something that the business can’t solve for them that must be hugely frustrating, which is obviously where they –
Susan Walsh [39:24]
: Or like I have heard instances where they’ve spent a year just fixing it themselves in Excel and by the time you fix that data, it’s out of date. So, that’s why I get calls in the early try, and try and get me in, and get me fixing things.
Jonny Dunning [39:48]: So, when you fix the data, how do you fix the process as well? Or do you fix process in the sense that if you’ve got, for example, teams over here entering information wrong and things misclassified and companies being added on with duplicate names and stuff like that, what do you do to kind of put stuff in place to try and not only fix the existing data but kind of solve the problem for them moving forward?
Susan Walsh [40:08]: I don’t get involved in that stage and quite often, I am the person where they just come to me and say, “I need this now. I just need the visibility.” I tried to get them to think longer term, “So let’s not just fix it for now, but let’s fix it for a longer-term so that you can get use out of this” and then I can train their teams to take over the classification and manage it, use this practice and which would hopefully, minimize the number of input errors.
The actual point of entry, I don’t get involved in that at all. And I don’t even think that procurement really necessarily have any involvement in that because that’s probably done by someone in finance for example if it's invoices.
Jonny Dunning [40:57]: Yes, or like CRM, it’s somebody in sales or it could be operations, but I guess if the underlying classification and the taxonomy and things like that, the structure and the consistency are sorted, then it’s only really going to affect a completely new data, completely new suppliers, completely new clients, all that sort of stuff where people could add additional errors to it.
Susan Walsh [41:19]: Yes. The theory would be that if you are not onboarding a lot of new suppliers every month or every quarter when you’re refreshing, maybe it’s like 60% you can auto-update the first time around, the next time it may be 80% and then you’re probably always maybe – or the closest you might get is like 90%, there’ll always be a better tolerance there or new descriptions that you’ve never seen before. So, it gets easier the more you do it, and again, it’s that maintenance, it’s that regular working with the data, getting to know it, it makes all the difference.
Jonny Dunning [42:02]: So, effectively once you’ve taken a company’s classic car-shaped data and kind of acid dip the chassis to get rid of all rust and redone the bodywork then maybe there’s a little bit of touching up here and there over as you progress, but the bulk of the work the structure is in place and it’s easier to then[42:23 inaudible] [cross talk] and maintain it.
Susan Walsh [42:24]: Yes, and then it’s up to them to polish it every week, check the wheels, check the tire pressure, all that stuff, all those bits that might not seem like a big thing, if you don’t do it, you’ll get called out.
Jonny Dunning [42:39]: And so, moving from the classic car analogy to a slightly different analogy. I see this in some ways as kind of like detective work because you’ve got to go in there and you’ve got to -- if you look at, for example, take an IR35 scenario where an organization is saying, okay, we’ve got 30 million spend that’s in services procurement tail spend and we don’t know what that is with your detective hat on or detective coat I should say. Sorry. How do you approach that? How do you go about that kind of collecting and sorting it out?
Susan Walsh [43:18]: Normally, what you’ll find especially if it’s a new dataset you might work on it the first day in their stuff that you just don’t know what it is, but after working with the data for a few more days that thing that you didn’t know on day one suddenly makes more sense because you get to know the data. I completely forgot the question. Sorry.
Jonny Dunning [43:42]: No.It’s just talking about it from that kind of detective point-of-view.
Susan Walsh [43:45]: Yes, and so a lot of the descriptions actually, lots of internal acronyms, figuring out what they mean and then knowing that if you find all of those you can classify it as a certain thing and you get a lot of internal references to things and even things like you might be working for a client who works with really large brands and so they’ve got expenses like, I’m meeting with another large company, so you have to be careful.
Say it was Adobe, they were doing business with Adobe, you have to be careful not to classify it as software because they had a meeting. It wasn’t software at all, it was a meeting. So, you always have to read between the lines, look at the context, look at the supplier; would they buy this product or service from this customer that I’m working for?
Tinder is the one that I used in the book and all the talks that I do and there’s a Tinder Corporation Limited here in the UK and that was in a dataset and I was thinking my client, there’s no way they are paying Tinder. No way. And of course, they weren’t because it’s an IT services company, it’s completely different. But it’s having that inquisitiveness, that knowledge that not just accepting things at face value and digging that little bit deeper.
Jonny Dunning [45:27]: What about just actually getting your hands on the data? Do you find there’s ever any kind of internal resistance? Do you find the stuff like, you say, well – Sorry?
Susan Walsh [45:37]: There’s low priority. I literally just got some data this week that I’ve been trying to get since December. They’re too busy on other projects to give me the data, so we’ll have to wait.
Jonny Dunning [45:50]: But do you also find hidden data treasure troves where people might have data stored locally on their desktop or it might be distributed within teams?Just the fact that you need to do almost like a data collection exercise, or is that going a step before?
Susan Walsh [46:10]: No, I trust them to give me the data and it could come from multiple sources. Ifit’s a global company, then they will have different source systems. But the type of data that I’m working with you wouldn’t expect it to be saved locally, it would be in a finance system or something like that.
Jonny Dunning [46:29]: Okay, so where companies are in situations where something we see a lot in services procurement is people actually literally using Excel spreadsheets to manage sometimes very significant amounts of spend and numbers of suppliers. How would you get involved in that sort of thing or is that kind of too immature really for you to be able to apply the principles you’re talking about?
Susan Walsh [46:56]: No, I still could classify it. I wouldn’t approach it in any different way like all the files come to me as Excel files. So, the only thing is I would just need to be aware of them, that would be the only difference. But you’d expect at least procurement to be talking to each other[47:13 inaudible].
Jonny Dunning [47:14]: Well, I think this is where the use of systems is important. So, if stuff is captured in a system, then obviously, you might have messy data, but at least you’ve got the data and at least you can do something with it. We definitely see instances where data is distributed and it isn’t in necessarily essential software system because if you look at, for example, services tail spend – pretty complicated, it's not really suited to any of the existing systems, it’s below the procurement threshold level of rigor and an assessment because it’s just not scalable to manage that manually and that sort of thing, it might sit nowhere in a worst-case scenario. What sort of advice would you give to companies that are in the situation where not only is the data not necessarily clean, but procurement teams aren’t actually capturing the potential data?
Susan Walsh [48:10]: Well, for a start, all data should be treated as equally, it doesn’t matter whether it’s a 1 billion pound/dollar client/supplier or hundreds buying because you don’t know that next year, they might be spending much more, significantly more. It’s just they've come onto your system at the end of a financial year or something so it looks like they've hardly spent anything with you. Just because there’s a financial value assigned to it don’t assume that that’s the value of the data. Read all exactly the same. You should really have it all in one place and it should always be backed up. Always. Always keeping an original copy.
Jonny Dunning [49:04]: Yes. And is that something you’ve come across before where companies have been working with a set of data and then lost data in some way or had those kinds of breakdowns?
Susan Walsh [48:14]: No. I work with really large organizations that have systems, so that’s fortunately never happened to me. I did work once with a smaller client who overruled a lot of the work I’d done to the point that it was unfixable, and so I made him fix it. Sometimes… I’m not doing this.
Jonny Dunning [49:39]: That’s fair enough.
Susan Walsh [49:39]: That was directory data, but yes, completely just messed it up beyond the point of any salvageable… and I guess they’re the ones that tend to have less security protocols in place for things like that or they don’t know and rather than kind of check with someone they’ll just have a go. If you’d just ask me first, I would have said right, we just need to just do this, and then it will be okay, instead of attacking it and then not being able to fix it.
Jonny Dunning [50:19]: So, in terms of when you’re doing that analysis that’s the discovery phase, the detective phase, obviously, you’re trying to understand the patterns and things like that, are there any kind of common data trails that you typically follow where you’re going through that process?
Susan Walsh [50:39]: It’s making sure that there’s not one type of spend sitting in two different classification buckets. So, sometimes IT consultancy sits under IT, sometimes it’s under professional services; it depends on who you’re speaking to, and some people have it in both. And so, you really want to have a taxonomy where that’s just not an option, you want to make each option clear and unique so that people can’t put it somewhere else.
And quite often, I’ll see people get subsistence food traveling confused with facilities catering as well, and while it might not seem like a massive issue, a level one, that’s a difference between facilities or travel. And if you roll up all those small little travel spends, that can add up to thousands or tens of thousands, so it’s never just one little piece of information on its own, it could roll up into a bigger problem or a bigger issue. So, it’s always good to be aware of that.
Jonny Dunning [51:53]: Yes, absolutely. One kind of last major area I wanted to just discuss with you was once you’ve done this work, now obviously you have your, okay, we’re all done, off you go, and what can be done with it and how can the business benefit? And how have you seen that kind of play out?
Susan Walsh [52:16]: So, a number of different things. First of all, the time saved in accessing that information or refreshing it, the time-save not having to fix so many mistakes, the opportunities for cost-savings that might never have been noticed before, spotting any kind of fraudulent activity or overspending with a supplier. Checking those things out, it’s more than just fixing the data and then maintaining it regularly keeps that level of effort down.
The more you maintain it the easier it becomes and actually, it becomes easier to spot things that are wrong because you’re familiar with the data. So, as I was saying, on day 1 of a file, I might not know something but by day five, I’m like, okay, that’s not right, that doesn’t look right, I can tell that should be something else.
Jonny Dunning [53:22]: But I’d imagine for procurement teams as well when you’ve tied it up their data and left it all gleaming and organized, that it’s probably easier for them to spot anomalies as well because they then have to make friends with the data.
Susan Walsh [53:34]: Yes, and I’ve stopped at the point of fixing the data so they can go off and do their own analytics and spend analysis. They should know their data to better than me so they should be able to interrogate it much better than I could, but I’m just lightening the load and doing the heavy lifting part at the start for them. I would always say to them, to hand it back to them, they should be responsible and manage their own data. They can do it better than anyone else, I’ll just help you get it to that stage.
Jonny Dunning [53:13]: I think this is really fascinating and specifically from my own particular area of interest around services procurement, I just think there’s so much data that can be unearthed, like services procurement is probably, in a lot of ways, is a bit behind in the sense that there’s a lot more data that can be unearthed and then probably it’s going to be messy data that needs to be cleaned up.
Susan Walsh [54:34]: Definitely.
Jonny Dunning [54:34]: But there’s a lot of data and information that companies can unearth and when you look at the use of services whether it’s in response to things like IR35 where companies are maybe reducing the use of contractors, maybe needing to get the work done in different ways, getting suppliers to do it under a statement of work, same applies to Brexit, more outsourcing because of that free movement of people, COVID made kind of deliverable-based outcome-based work more acceptable.
So, I think it’s only going to be a growing area and I think it’s probably going to grow in the tail spend particularly. I think there are huge opportunities there for organizations to capture that within a dedicated system and then do some cool stuff with the data. Just to sort of wrap things up, obviously, you’ve had a very busy year, I know your business has grown a lot over the last year as well.
Susan Walsh [55:22]: It’s just doing about.
Jonny Dunning [55:24]: How much has it grown?
Susan Walsh [55:26]:500%.
Jonny Dunning [55:28]: Obviously, you’re doing something right. So, basically, you’ve had this period of real kind of dynamic change in your business. You’ve written the book, so obviously, that’s going to come out in the next six months at some point. What do you envisage for the next 12 months of your business and what do you see as a kind of major things going on in the market that might impact that?
Susan Walsh [55:54]: So, for me, I’m not quite at stability mode yet, the work still fluctuates. It is project-based and maybe that will never change. But if I can get more project work in, then those dips will be fewer. So, I’d like to grow the team a little bit more and maybe offer additional services. In terms of how things are going to go, I think that there’s going to be even more cost-savings required by businesses coming down the line.
So, it’s going to be more important than ever to check that data and not even that but just in terms of sales and marketing teams targeting the right customers, you have to have that kind of data right as well, has to be accurate in date with the right email address or postal address. You need to know your manufacturing plans and you need data to be able to look at to build forecasts, you need to look at what’s happened previously and it’s a bit of an unknown for everybody, but by having data it will help guide you. And I think that’s only going to become stronger.
Jonny Dunning [57:12]: Yes, I totally agree with that. I think what you’re doing is fantastic, really pleased for you with how well everything’s going and with the book and everything like that. And I really appreciate you taking the time to come and have a chat. I’ve really enjoyed that.
Susan Walsh [57:24]: No, it’s been great.
Jonny Dunning [57:26]: Some great insights there and I think you’ve got a very unique perspective.
Susan Walsh [57:29]: Yes, but I think you’re onto something with the whole professional services thing as well. There’s a lot of hidden -- there’s a lot of like my mates in there, just does about work every now and again, but it’s just investing every month. There’s probably a lot of that going on, and how will you know if you don’t have the data?
Jonny Dunning [57:49]: I think you’re absolutely right, if you can capture that data, then you can make sure the data is valuable and you can do all the valuable things with it. And where businesses are spending large amounts of money and services procurement is somewhere between a trillion and 20 trillion annual spend globally.
Susan Walsh [58:06]: That is a lot.
Jonny Dunning [58:07]: Yes, something like that. It deserves some specific scrutiny. But yes, really love what you’re doing, and thanks so much for taking the time. Really appreciate you coming on.
Susan Walsh [58:18]: Thanks for having me. Thanks a lot.
Jonny Dunning [58:21]:[inaudible].Bye.
Susan Walsh [58:21]: Bye.