So for email marketing, I initially built a custom coze agent for automatic email workflows. But after some time, I decided to transition to an existing solution because they offer more mature systems, integrated functions, and a one-stop approach. Through extensive research, I came across Instantly and have been building my email campaigns on their platform ever since.
First step: set up your email account. There are numerous tutorials available, and it’s critical to follow each step precisely—missing a step can hurt email deliverability. Instantly has a pre-warm function, which is a game-changer: using a personal email often leads to frequent blocking (affecting personal use), and a brand-new corporate email might face reliability issues since it hasn’t been validated by email providers. Instantly’s warm-up feature handles this perfectly.
Next, move to campaign setup—this is where the magic happens. I recommend tagging each campaign: for example, one tag for subject lines with dynamic variables, another for a general non-personal template (like “version two”). This lets you run A/B tests across campaigns. Start by uploading leads (I typically import from a Google Sheet; authorization is straightforward), then customize variable types. For the email sending phase: define the subject line and HTML template. I use Claude Plus Cursor to build flyers, but remember to stick to email-compatible CSS—fancy CSS might not work on Instantly (some email providers block complex formats). Always preview your email before sending! Also please run the email body through an LLM to scan for spammy text—overusing spam triggers can get your corporate email blocked by providers like Microsoft (a major red flag).
Lastly, schedule your sends: set a minimum 5-minute gap between emails to avoid looking like a bot, cap daily send limits, and enable link tracking. This records open rates, click rates, and replies in real-time analytics—super handy for monitoring performance. I also use UTM tracking (via a UTM link builder) and a Python script to clean leads, add unique user IDs (which allows me to later filter out bot traffic). With UTM tags embedded in links, you can pull data into Microsoft Clarity or Google Analytics to compare campaign performance and validate A/B test assumptions.
