Why publish your own numbers?

Because an automation agency that will not show its own automation is asking you to take a lot on faith. These numbers are measured from our own logs and campaign dashboards. Where something is an estimate, we label it.

Would the same automations work for my business?

The categories transfer: outbound, content, lead handling, reporting. The implementations should not be copied blind, which is why every engagement starts with a $500 audit of your workflows rather than a template.

What do you still do manually?

Sales calls, pricing decisions, final review of anything a customer sees, and every strategic call. Automation does the repetitive middle. The judgment stays human, on purpose.

Build Log: We Run DappaSol on Our Own Automations. Here Are the Numbers.

DappaSol runs its own operations on the automations it sells: automated outbound that sent 1,636 cold emails for roughly $49, a personalized pipeline behind 62 hand-audited outreach emails across 9 niches, a content engine maintaining 60+ guides as Markdown, and self-hosted lead capture. This build log publishes the real numbers, including what broke.

I am Ishan, I run DappaSol, and we sell ops automation. The obvious question a buyer should ask any automation agency is: fine, show me yours. This is that page. Real numbers, from our own logs, including the part where an automation misfired.

Outbound: 1,636 emails, about $49, one lesson

We run an automated cold outbound system: prospect sourcing, enrichment, sequencing and sending, on a capped budget. One batch: 1,636 cold emails for roughly $49 in send spend, which produced about 15 replies and 8 warm conversations. Those are small numbers and I am publishing them anyway, because they are real, and because the interesting part is what happened next: we paused the whole thing.

Why: new campaigns in the tool defaulted to a geography we never chose, and unreviewed sends went out before we caught it. The fix became policy: a $15/day spend cap on the account and a geo check before any campaign starts. That incident is the most honest automation lesson on this page: the automation did exactly what it was configured to do, and the configuration was the bug. Design the guardrails, not just the workflow.

Alongside the volume machine we run the opposite play: 62 outreach emails across 9 SMB niches, each one built on an automated audit of the prospect’s actual website, with the findings written into the email by hand. Automation does the research grunt work; a human writes the sentence that gets replied to.

Content: this library is the machine’s output

The page you are reading is part of the system. Our guides live as Markdown files in a content collection, 60+ of them, produced through an automated pipeline: research, competitor reading, drafting, internal linking, schema, with human review before anything publishes. Supporting page sets ship the same way, like our answers set (5 pages) and fix set (6 pages). When we doubled down on the AI-security cluster, five full guides shipped in one working session.

The honest caveat: the pipeline drafts, it does not decide. Topics, claims, prices and every ranking judgment are human calls. Automated publishing without that layer is how the internet filled up with sludge, and we would rather ship fewer pages than join it.

Lead capture and the unglamorous middle

Every lead form on this site posts to a self-hosted endpoint that appends to a JSONL log and emails us. No CRM subscription, no third-party form SaaS, nothing that breaks when a vendor changes plans. Boring, owned, reliable: which is the whole automation thesis in one line.

The same philosophy runs through the rest of the middle: an ad-creative pipeline that composites captions and end cards onto 9:16 video locally (we have not scaled paid spend yet, so no ad performance numbers to publish; when there are real ones, they will be here), and a capture pipeline that turns our WebGL builds into the autoplay loop videos on the work page.

What stayed human

Sales calls. Pricing. Final QA on anything a client sees. Deciding what to build next. The pattern across everything above: automation owns the repetitive middle, humans own the ends, the judgment and the relationship. That is also exactly how we scope client systems.

If you want this for your business

The same categories (outbound, content, lead handling, reporting) apply to most businesses doing repetitive ops. Start with what business automation costs, or the honest DIY route if your volume is small. When you want it built, the $500 audit maps your workflows and ranks them by ROI, and builds start at $6,500. Book a 15-minute intro call.