Any D2C organisation running performance marketing at scale will eventually hit the same wall.
You have optimised your search campaigns. Brand and non-brand keywords are well covered. You are scaling Facebook. You are squeezing efficiency in every turn, impression to click, visit to lead, lead to sale. And then you realize that search demand for your category is not growing fast enough to give you the headroom you need. The channel has a ceiling.
You are confronted with two options. Find new channels for fresh leads. Or find a way to activate the leads you already have but never converted.
Most large performance marketing operations are sitting on millions of unconverted and dormant leads. People who dropped a lead in a landing page. People who started a quote. People who spoke to a telecaller and then dropped off. Each of these represents real intent at some point in time. Most organisations have not figured out how to reactivate them at scale.
That is the problem we set out to solve at Star Health. The IVR and WhatsApp automation channel was our answer to it.
The asset hiding in plain sight
When you look at a dormant lead database, the instinct might be to see a graveyard. These people did not convert. leads went cold. Move on.
But that is not the right perspective. These are people who knew your brand and showed enough interest to engage at some point. Maybe, life got in the way or the timing was not right, or the right follow-up never reached them or the picth by your telecaller was bad. That is different from someone who does not know your brand.
The question is not whether to reactivate them. The question is how to do it in a way that is relevant and does not feel like spam.
How the channel works
Consent before everything. Both IVR and WhatsApp were used to generate fresh consent before any structured outreach begins including calling. The IVR asks the customer to confirm if they want to hear from the Brand. That single step filters for intent immediately. WhatsApp works the same way. Your IVR intent is soldified via WhatsApp opt-in. Getting fresh consent is also how you stay on the right side of regulation and trust.
WhatsApp is a journey, not a blast. Once you have consent on WhatsApp, you can build a sequence that plays out over days or weeks depending on how the customer responds. IVR is a single interaction. WhatsApp is a conversation that unfolds. In WhatsApp, you are not broadcasting at scale, you are walking each lead through something that feels relevant.
Assumptions over certainty. You rarely know exactly where a dormant lead is in their life. So you work with reasonable assumptions and write messages that would hold up across the likely scenarios. Someone who enquired but never bought: assume they are still without cover and message accordingly. Someone who may have gone to a competitor: message with porting options. If your telecallers captured good disposition data at the time of the original inquiry, you have more to go on and can tailor more precisely. Most of the time you are making educated guesses, and that is fine as long as the guesses are grounded.
Segmentation from Data signals. Data signals determine what angle to take. A lead who flagged financial issues with a low affluence score gets a message about EMI or BNPL options. A high affluence lead from a metro gets a message about higher sum insured and more comprehensive cover. And any lead who has not been in touch for a while should hear about what has changed since they last spoke to you: a new product, a new add-on, a price point that did not exist when they first enquired. That is often reason enough to come back. A lead that had sufficient amount of talktime, but has not decided yet needs a reassuring testimonial from an existing customer or a hospital partner from their region.
The metrics that mattered
The metrics are the same ones you would use to evaluate any performance marketing channel. That comparability is important: it lets you put this channel next to search or paid social and make an honest call on where to invest.
- Lead connectivity (%): what percentage of leads in the sequence are you actually reaching
- Lead to Qualified (%): of those reached, how many express genuine interest and move forward
- Lead to Sale conversion (%): end-to-end conversion from lead to policy
- CAC: cost to acquire a customer through this channel versus others
- ROI: revenue generated relative to the cost of running the channel
Of course, we tracked channel level metrics too viz., Read rate, response rate, Opt-outs, delivery to lead, so that the messages, segments, delivery timings etc. can be optimized to get more of them back in the sales funnel.
Over time, the channel grew into one of the top contributors to D2C fresh sales at a blended CAC that worked.
What the playbook actually requires
The technology is not the hard part. Most capable CRMs or Marketing Automation Platforms (MAPs) can run drip journeys. WhatsApp Business API is accessible. Anyone can set this up technically. The hard part is three things.
Data discipline. The messaging sequences only work if your lifecycle states are accurate. CRM hygiene has to be real. Garbage in, Garbage out i.e., you may end up sending irrelvant messages.
Feedback loops. You need a team who can read weekly performance data and translate it into messaging changes, sequence adjustments, segment refinements. This is not a set it once and forget system. It is a continuos process. Without the feedback loop, performance decays.
Cross-functional trust. Marketing automation that feeds a sales team only works if the sales team believes in it. We spent real time showing agents which warmed leads were converting better and why. Adoption follows understanding, not instruction.
The bigger point
This is not an insurance specific playbook. Any organisation running performance marketing at scale has the same asset sitting in its database: years of unconverted intent.
The growth question most teams focus on is how to get more leads in. The more interesting question is what you are doing with the leads you already have. At scale, the difference between a meaningful reactivation rate and a poor one on a dormant database is a significant revenue number. That is a channel worth building.
This essay describes work done at Star Health Insurance. Numbers referenced are directional and do not represent official company figures.