Siddharth named two conditions for success: that our search and AI visibility beats incumbents on substance, and that owners actually send us their evidence. This memo takes each in turn, prices the first against what Indian directories charge, and says which one carries the risk.
Ranking runs as a sequence, and freshness sits near the end of it.
Freshness is a tiebreaker within a set we have already qualified for. It rescues nothing that failed step three. A new domain publishing weekly stays below JustDial on “solar installers in Surat”, because JustDial carries a decade of links and entity weight. Recency multiplies an authority base, and ours starts near zero.
Citation by an AI assistant behaves differently from a blue link in three ways, and all three favour us.
So the realistic path runs through the long-tail data questions first, accumulating citations and authority from there, then moving up. Recency plus proprietary data plus long-tail specificity plus time, with recency the weakest of the four.
Anyone can copy the freshness mechanic. The data supply resists copying, because it arrives through customer relationships. Lead with that.
| Offer | What the business gets | ₹ / year |
|---|---|---|
| JustDial, entry | Paid listing with photos; leads passed on (reported) | ~20,000 |
| JustDial, national | Higher visibility tier (reported) | ~40,000 |
| IndiaMART — MDC | Mini Dynamic Catalog, the base package (reported) | ~28,000 |
| IndiaMART — Maximiser | Lead package, around 30 BuyLeads a week (reported) | ~60,000 |
| IndiaMART — TrustSEAL | Verification badge (reported) | ~45,000 |
| IndiaMART — IM Leader | “Leading Supplier” add-on, sits on top of a base package (reported) | ~45,000 |
| IndiaMART, realistic stack | Base package plus IM Leader, as an SMB would actually buy it (reported) | 70,000–1,05,000 |
| Fairview today | Managed presence across every surface a buyer or assistant checks | ~2,16,000 |
On every figure above. Neither JustDial nor IndiaMART publishes a rate card. Reps quote per category and per city, and multi-year commitments are negotiable, so each number is a third-party reported band. We have confirmed none of them. Older material cites materially different tiers, which suggests the naming and the pricing have both moved over the years.
On the Fairview figure. ₹2,16,000 a year is derived, calculated as ₹18,000 a month across twelve months. That monthly figure is itself a blended average across the book (₹5.6L ÷ 31 customers ≈ ₹18,065), so real plans presumably sit either side of it.
We stand roughly ten times above JustDial’s entry package and two to three times above a realistic IndiaMART stack. A like-for-like price argument against a directory is unwinnable, so the anchor has to be the agency retainer, where ₹18,000 a month is ordinary or cheap for a website, content, social and a Google Business Profile together. The directory spend becomes redundant over time, and it never was the comparison.
A ₹1.9L solar job at roughly fifteen per cent margin is about ₹28,500 gross. One extra job every two months covers ₹18,000 a month. That is an easy conversation to have, and it means a solar EPC bears far more than a salon can. Price by the ticket size of the trade.
“Five thousand mixed SMBs” hurts us twice. It thins the corpus below the density citation needs, and it drags average willingness-to-pay down. Density inside high-ticket trades helps us twice, on both of the same axes. Our concentration in solar is an asset on both counts.
At ₹18,000 a month we will never reach the density a citation flywheel needs inside any single trade-and-city cell. JustDial solved exactly this problem: free listings build the corpus, paid listings monetise it. A cheap or free verified-profile tier is a flywheel requirement, and it happens to be the thing our SDG 10 slide already promises. Worth designing deliberately.
Forwarding evidence to a bot is something the owner has to start doing, and new habits that need owner initiative decay hard. Our own comparison slide says a DIY builder loses the owner by week two. The same mechanism applies to us. Our ask is much smaller, which is real and matters, though it stays above zero.
If owners stop sending, we source the evidence another way. Bulk-import the photos already sitting on their phone, integrate with their invoicing, or have our team chase it. All of those work, and all of them convert software margins into services margins. The downside is margin compression, which is worth knowing, because it changes how much we should spend to avoid it.
Bet one has a known answer: yes, with unique data and twelve to twenty-four months of authority. It is a slog with a method. Bet two is a behavioural unknown, and no amount of engineering guarantees that a person sends a photo.
The two are also linked in one direction. No evidence means no fresh data, which means no unique pages and no citations. The search bet sits downstream of the habit bet, so one gates the other.
We have 31 customers. Run it by hand. Ask for evidence over WhatsApp manually, and measure the unprompted-send rate week by week. The decay curve from week one to week twelve is the single most important number in the business, and we can hold it in a quarter without writing code.
That number would also settle the question underneath the deck: whether “zero behaviour change” is a claim we can make on stage.