Strategy first, software last. Yes, even AI.
Clients often come to us to ask about technology decisions. What platform should I use as a CMS? CDP? Paywall technology? ESP? CRM? CRO? Marketing automation? Web analytics? And now, generative AI. And we always ask the same thing: what are you trying to achieve? Why are you even in the market for new software? This framework feels especially important amongst all the AI hype. McKinsey states this succinctly, “While there is merit to getting started fast, building a basic business case first will help companies better navigate their generative AI journeys.”
Your goal is the foundation for your requirements
Many organizations treat their tech stack like a fashion purchase: they want the shiny new thing that everyone else has. But software is a tool, and tools are used to build things. Instead of asking: what software is my competitor using? Ask yourself broadly: what do I need to build to be successful in the market? Then go from there.
Start with your goal, then back into your requirements, then narrow down your software options.
This means that before you start evaluating software options, identify the pain point you’re trying to solve for and/or goal you’re trying to achieve. Put some concrete numbers around this, if you can, like “decrease customer support wait times by 10%” or “increase conversion rates from email by 2x.”
Once you have the business case for your software investment, it’s now time to move into requirements building. This is where you build your list of must-haves as well as a wishlist. Requirements can include integrations with existing systems, privacy concerns, usability, and implementation timelines.
For example, let’s say you want to improve cross selling and upselling on your ecommerce site, as well as increase the efficiency of your customer support team. You need a tool that integrates with your existing content management system, ecommerce platform, and customer support system. Let’s say you use WordPress, Shopify, and Zendesk. Your metrics for success would be an increase in cross selling and upselling conversion rates, a decrease in customer support wait times, and no change or an increase in customer support satisfaction surveys. Those are your requirements. Now you can go demo a selection of generative AI chatbots that meet those requirements.
Humans implement software
Assuming you’ve started with a strategy and selected the appropriate software to build it, you now need actual humans to use the software. Yes, we’ve all read the headlines about how AI is coming for all of our jobs. That’s not what this article is about, though. Right now, if you’re going to roll out a new software solution, you need someone to be in charge of overall implementation and driving adoption. And if you want it to be truly successful, implementation can’t be a project you pile on a colleague whose plate is already full. This might mean that you need to bring in a specialist on a contract basis, or align OKRs to your software adoption goals. Whatever you decide to do organizationally, getting critical new software up-and-running shouldn’t be a side hustle.
Buy for now, plan for later
When evaluating software options, it’s easy to be dazzled by high-tech, very fancy features. But don’t buy a jet plane if what you really need is a midsize sedan with good gas mileage. Select the software you can actually use, right now. That said, you don’t want to spend months selecting and implementing a software solution that won’t meet your business needs within a year. So it is a good idea to choose a solution that has room for growth. Many B2B SaaS companies are on this: they also want to grow with their customers. This is the reason behind the ubiquitous three-tier pricing structure: single user, SMB, enterprise.
Digital publishing client example
Among our clients in digital publishing, we frequently hear this question posed around paywall technology options — from home-grown to enterprise-level options. We’re platform agnostic at Sterling Woods, and we’ve seen clients achieve success with a wide range of tools. The key to success is to pick the right tool for your model, which means starting with a comprehensive review of your website analytics. Once you’ve determined your paywall strategy–and that strategy should include a test-and-learn plan–then you can proceed with tool selection.
How Sterling Woods uses machine learning
Here’s our own approach. Sterling Woods uses machine learning in our Scout X service to quickly and effectively analyze our clients’ data to find high-value customer segments. We built the ML tool so that we could spend more time helping clients and less time mired in data analysis. The specific pain point we were trying to solve was that we needed to get key learnings about customer segments to our clients as quickly as possible, so that clients could strike while the iron was hot.