“My Data Sucks”: The Hard Reality Check Most Procurement Teams Need Right Now

Everyone wants to talk about AI. But when it comes to the “unsexy” work of fixing the data that feeds it, the room suddenly gets very quiet.

In this week’s episode of Proc & Roll, hosts Conrad Smith, Zachary Bachir, and Natasha Gurevich tackled the elephant in the room: Data Readiness.

Is your organization actually ready for predictive analytics? Or are you just stuck in a cycle of expensive “cleanup projects” that never last?

Here are four hard truths from the episode that every procurement leader needs to hear.

1. The “Bedroom Tidying” Trap

We’ve all been there: You hire a consultant, pay them to clean up your supplier master, and pat yourself on the back. Six months later, it’s a mess again.

Zachary Bachir calls this the “Bedroom Tidying” fallacy. “You’ve just done some bedroom tidying up, but you’re not keeping your bedroom clean every single day,” he argues.

The solution isn’t another cleanup project. It’s fixing the operating model. If your process allows users to create duplicate suppliers or enter bad data without checks, no amount of cleaning will save you. As Zach puts it: “No new data unless it’s absolutely necessary.”

2. The Sentence That Kills Strategy

Natasha Gurevich pinpointed the exact moment a procurement strategy falls apart. It’s not when the budget is cut—it’s when you present your plan, and a stakeholder looks at your slide and says: “I disagree with your numbers.”

Without a “Single Version of the Truth” that Finance, Legal, and Procurement all agree on, you have no foundation. Natasha warns that companies miss revenue targets precisely because they are building forecasts on disconnected data that doesn’t reflect reality.

3. You Probably Don’t Need a Data Scientist (Yet)

With the AI boom, there is a rush to hire PhD Data Scientists. But is that who you really need?

Natasha shared a cautionary tale of a PhD hired to do advanced predictive analytics, only to leave in frustration because the organization’s data maturity was at a “second grade” level. “It’s like if you take a second grader and you put them in geometry class,” she explains .

The verdict? Hire a Data Analyst or a process expert first. You need someone who understands SQL, tables, and data architecture to build the foundation before you bring in the scientist to do the fancy math .

4. Stay in Love with the Data

As leaders rise through the ranks, they stop running reports and start looking at dashboards. Conrad Smith argues this is a mistake.

His advice to every aspiring leader? “Stay in love with the data.” . Even if you aren’t the one running the queries anymore, you cannot afford to lose the ability to speak the language of data. In a world of AI and “digital first” strategies, the leaders who win will be the ones who understand exactly what lies beneath the dashboard.

Are you ready to stop tidying up and start fixing the process?

Watch now or read the transcript below.

Transcript: Proc-N-Roll | “You’re Only as Good as Your Last Failure”: Former Meta S2P Leader Talks Credibility

Conrad Smith: Well, what’s on the docket for today? What’s on the calendar here?

Zachary Bachir: Well, something completely different. We’re going to discuss data today and data readiness for AI. How do you make sure your procurement organization is set up to take advantage of AI from a data perspective?

Zachary Bachir: I’m in two camps here. On the one hand, yes, you absolutely need to look at your data and clean it. But I really don’t see a lot of organizations doing that systematically. I think it takes a bit of investment and it’s very hard to build a business case around data… It’s not always the most interesting or pressing topic like, “Hey, procurement wants to clean up their data because it will do X, Y, and Z.”

Zachary Bachir: I think it needs to go hand in hand with the AI strategy… I’m on the fence around how much and how fast to go on the data question.

Conrad Smith: Zach, data quality was a topic and measured even before AI was talked about. As you go around doing these engagements… how often do you see data quality as a topic as part of the project?

Zachary Bachir: Like every project data quality is every single one… But I just see a lot of these programs, the data doesn’t really get to… it’s not the strategic enabler in many cases. There are many things that the transformation needs to juggle. Data is just one of them.

Natasha Gurevich: I am initiating a lot of conversations with a lot of companies because their data sucks and they either do not know about it or they do not know how to address it.

Natasha Gurevich: They cannot have an accurate evaluation. What’s the size of the prize for them? …knowing where the money goes to, who the money is spent with… When you do not have accurate data, I don’t think you can plan accurately.

Conrad Smith: Zach, you said data is not the strategic enabler. I think I would strongly disagree with you. I do think it’s the strategic enabler, but it’s probably not the top line of the ROI. Nobody’s going in and convincing the leadership, “Hey, we need to do this big transformation because we have bad data.”

Zachary Bachir: When I said it’s not a strategic enabler, what I mean is these projects that companies do to clean up the data as part of some kind of transformation… These are really one-off activities. Where you really get the strategic enablement is cleaning up your processes for capturing data and maintaining it over time.

Zachary Bachir: If you’re having to clean your data every year, then you’re doing something wrong.

Conrad Smith: When I arrived at Adobe… we were paying GEP every quarter to go through a cleanup process. We would give them a file, they would do stuff on the file and give it back… It just always shocked me… in like five minutes, you don’t even have the total spend anymore.

Conrad Smith: If you don’t build, as Zach’s talking about, the model so that your process creates clean data… You’ve got to go all the way back to the root cause and say, “What are we doing here that’s creating bad data?”

Zachary Bachir: The larger companies obviously investing in data lakes and data warehouses… The problem with data lakes is data lake does not belong to procurement. You’re just one set of data in a vast number of different operations across the business.

Zachary Bachir: This idea of like, “I’ll do a one-off little investment, clean up my data”… No, no, you’ve just done some bedroom tidying up, but you’re not keeping your bedroom clean every single day of the year.

Natasha Gurevich: I think every procurement professional had been in a conversation where you present the data and someone in the room says, “I disagree.” And this is when your heart sinks… because someone disagrees with a number… your entire strategy is just falling apart.

Natasha Gurevich: Every department within the large enterprise has their own data team… and they use different data points and there are multiple versions of truth.

Zachary Bachir: I would have three bits of advice for companies looking at data.

1. Know what data is important. Link it back to whatever your strategic objectives are for that year.
2. Don’t introduce new data unless it’s absolutely necessary. Don’t let them log a supplier name anywhere until they’ve absolutely checked that the supplier doesn’t exist already somewhere else.
3. Document the data architecture. A clear taxonomy of your data and what systems have which data points.

Conrad Smith: Would you hire a data scientist or would you hire a process expert?

Zachary Bachir: I don’t think you need data science right now in procurement… You need someone who’s a data analyst… who deals with large quantities of data and creating structured SQL tables… usually that’s not gonna be someone from procurement.

Natasha Gurevich: I’d say that the need to bring a data scientist depends on maturity… We had a person who had PhD in data and was hired specifically to do predictive analytics… The person had been extremely frustrated… It’s like, if you take a second grader and you put them in geometry class and you say, “You figure it out.”

Conrad Smith: Now more than ever, you need to be in love with data. The AI first or the digital first… think they’re kind of taking us further and further away from really being in love with the raw data. Don’t stop speaking that language.

This transcript has been edited for clarity while maintaining all substantive content