will warnock

Data Strategy & AI

What I've found in my time working with New Zealand companies is that there comes a time when business data begins to matter. The crossover point into this mattering stage is two-fold. Data-related challenges and opportunities start to reveal themselves throughout your business.

If your business is anywhere near this stage, the below will be familiar and enticing to you.

• Switching operational systems will involve a long and costly migration process
• Building business reports is hard and/or takes too long
• Data sources for reporting feel disjointed, making it difficult to generate reports with business value
• Wondering why AI is not assisting your business

• Business performance at the tip of your fingers
• Identify sales trends or seasonal impacts
• Plan for inventory replenishment
• Integrate systems for automation of data entry
• Leverage new AI technology with actual business value

Whether you've found yourself keen on mitigating data challenges, looking to leverage new opportunities, or a combination of both, what you're after is a data strategy. A good data strategy will allow you to derive significant value from the business data you already have and provide agility to adapt to a rapidly changing data landscape.

When I say data strategy, I'm not referring to the definition you'd find in a textbook. Rather, I mean have you asked yourself the question, how is my business data being stored and how do I retrieve that data? For many of the companies that I have worked with, the answer to the first part of this question is that their data is distributed across separate Inventory, Ecommerce, Accounting and CRM systems, all siloed by function. Given this, the task of retrieving data out of these systems is arduous and time-consuming, frequently consisting of running separate reports in each system, downloading large Excel files and then stitching these sources together with a barrage of anything-but-best-practice-formulas, lookups and highlighted cells. In many businesses, this mess of a final Excel file is referred to as the "Holy Bible" of the business due to how strongly the staff come to depend on it.¹

Now a couple of years ago I'd tell you that the solution to this is simple: pipe your key data sources into a Business Intelligence/Analytics tool, build some reports, and call it a day. While this is where we want to end up, it's not how we get there. There are some challenges with storing all of your business data inside an analytics tool that are frequently not considered:

• What if you want to switch reporting tools?
• What if the software vendor increases its prices by X% every year?
• What if the vendor gets acquired and the service stops receiving updates, is merged with another product, or turned off entirely?
• What if the tool fails to keep up with new features offered by competitors, like AI functionality?

In each of these cases, there's significant time, cost and effort involved with switching analytics tools. You need to re-pipe all the data from your existing systems, clean the data, transform it and establish any relationships between tables so you can re-build your custom, charts, reports and dashboards.

Introducing the data warehouse. A central database that will serve as your one-stop shop for relevant business data. Big global enterprises have been using these for years, but with the opportunities associated with AI for business applications, any NZ company where their data matters should be utilising a data warehouse. Here are the perks:

• Greater control and sovereignty of your data
• Increased security and permissions controls
• Data cleaning and manipulation can be hidden from the final user to reduce noise within the analytics tool
• Increased flexibility to use the data downstream (alternate reporting tools, raw Excel/CSV exports)
• Retain historical records for future comparison
• Makes your business AI ready

How it works is that by instead of integrating your data directly to your analytics tool, it is synchronised to a centralised database either in the cloud (AWS, Azure, GCP) or on-premise (ok boomer). As an example, accounting data from Xero, inventory data from Cin7 and CRM data from Hubspot could all be integrated to flow into the data warehouse and be stored in their raw state. Then, with some transformations this data can all be joined together and structured into flat views like "Sales by Product", "Sales by Customer", "Aging Debtors" or "Marketing campaign opening rates". These views can then be synchronised to an analytics tool of your choosing. By centralising your data sources here, you can be confident that it is correct and consistent throughout your organisation whilst keeping ownership of it and having flexibility to use it how you wish. Some example scenarios that become possible through a data warehousing approach include:

• Observe how many days individual customers are taking to pay you, and how this compares to their payment terms and previous payments
• Track performance and return on investment of marketing campaigns by analysing spend on Meta/Google Ads against sales data from the same time period
• Create draft document proposals that are addressed to new prospects in CRM that use previously created documents as working templates
• Create a knowledge base of company processes and help articles for new employees being onboarded through an AI that sits across your data warehouse
• Ask an AI which of your products have the strongest margin based on average sales prices from your Ecommerce data and average unit cost from actual bills paid to suppliers in your Accounting system

Like most new technology advancements, speculators are likely overestimating what AI can do in the next two years, but underestimating its impact over the next decade. No one can predict the future of how this will shake out, but you can become ready for AI by adopting a data strategy for your business. Any AI application is only going to be as strong as the data you can provide to it, and if you're considerate and future thinking, then you'll be able to generate strong business value from AI in the future as it matures and evolves.