East London Lets is a five-year-old independent lettings agent covering E5, E8 and E9. Adesh runs it solo with one part-time admin. The previous site sat on a "letting agent SaaS" platform that bundled CRM, listings sync and a templated website at £140 a month. The CRM worked; the listings sync worked; the website had been invisible to Google for two years. Search Console showed zero indexed pages and zero clicks — a textbook example of a "wide but shallow" template platform tripping the Helpful Content classifier.
The brief
A custom site that pulled the property listings from the existing PropertyData feed (Adesh did not want to abandon the CRM), rendered each property with proper RealEstateListing schema, ranked for the local-pack queries Adesh cared about, captured valuation enquiries straight into his CRM via webhook, and presented the brand with the credibility of a high-street agent rather than a faceless aggregator. Five fixed pages plus the dynamic property routes.
What the audit found
The agency-template platform had three structural problems. (1) Every "letting agent" subscriber was sharing the same boilerplate copy — Google was correctly identifying it as duplicate-content boilerplate and de-indexing across the network. (2) The schema emitted was generic Organization with no RealEstateListing entries per property — Google had no way to surface specific properties in the rich-results listings. (3) The page templates were render-blocking enough to fail Core Web Vitals on every property page. The platform was a working tool for the CRM but a structural ranking hazard for the website layer.
What we built between 9 AM and 5 PM
A five-page Next.js site plus a dynamic property route that pulled listings from PropertyData via a typed API client and cached at build time (with on-demand revalidation when listings update). The home page led with the value proposition and the hyper-local copy section naming the streets and developments Adesh specialises in (Clapton Square, Hackney Downs, Lower Clapton Road, Mare Street). The "valuation" page held a multi-step form that fed the existing CRM via webhook with the right lead-source tagging. The "areas we cover" page broke the coverage area into E5, E8 and E9 sub-pages with neighbourhood-specific copy. The "about" page covered Adesh's background. The dynamic property pages each carried full RealEstateListing schema with the floor area, the EPC rating, the council tax band, the deposit, the available-from date, and three photos.
The PropertyData integration
PropertyData (the existing CRM) exposed a JSON feed of active listings. The new site polled the feed at build time and on a scheduled cron via Vercel's revalidation hook, generating static pages for each listing. The 14-day cron handled the long-tail of listings that came on and off market without needing the agent to do anything in the new system. When a property went under offer, the page updated within an hour. When a property let, the page redirected to a category landing rather than 404 — preserving any inbound link equity.
The schema and SEO layer
Per-property RealEstateListing entities with the address, the price, the floor area, the EPC rating, the council tax band, the available-from date, and three photos. RealEstateAgent schema for the agency itself with the address, the areaServed (E5, E8, E9 enumerated), the aggregateRating from Google reviews, and the membership of the relevant industry bodies (ARLA Propertymark, The Property Ombudsman). Every page indexable, every page in the sitemap, every page with hreflang en-GB.
The launch and indexing
DNS swap at 4:50 PM. Search Console URL inspection at 5:00. By 8:30 the next morning Google had crawled and indexed the homepage and three property pages. Inside 48 hours, all five fixed pages and 23 of the 26 active property pages were indexed (the remaining three indexed within the week). The previous platform's zero-indexed-pages baseline had been replaced with 28 indexed pages inside two working days.
Week one valuation enquiries
Four landlord valuation enquiries landed via the multi-step form in the first week — three for properties in E5 and one for E9. Adesh closed two of the four into instructed listings within the month. The valuation form's lead quality scored materially higher than the previous platform's contact-form leads (the multi-step structure self-filtered casual enquirers), and the time from enquiry to instruction averaged 22 days against the previous baseline of 41.
The ranking trajectory
Week 2: indexed for "letting agent E5" and "letting agent Hackney" at positions 18 and 22. Week 4: positions 11 and 9. Month 2: positions 7 and 5. Month 3: positions 6 and 4 — the latter ahead of one of the national portals on the long-tail combination "letting agent Hackney E5". The trajectory has been stable since; the agency now competes head-to-head with the national portals on Adesh's specific micro-geography, which is the only geography that matters for an independent.
What changed about how Adesh runs the business
Pre-launch, valuation enquiries came almost entirely from existing tenant referrals (which is a slow channel) and the occasional Rightmove lead (which is expensive and not exclusive). Post-launch, the website is the largest inbound channel by volume. Adesh has reallocated the £140-a-month agency-template subscription into a Google Ads campaign for "letting agent hackney" and adjacent queries; the campaign produces an additional six to eight valuation enquiries a month at a cost-per-acquisition of roughly £18, comfortably inside the firm's £45 target for valuation work.
The compliance and trust layer
Independent lettings agents trade on local credibility. The site reinforces it: the ARLA Propertymark badge and registration number in the footer; The Property Ombudsman membership and complaints route prominently linked; the agent's Client Money Protection scheme details on the about page; the deposit protection scheme details on every property page; and the full GDPR-aligned privacy notice covering the prospect-to-tenant pipeline. Each of those is a small E-E-A-T signal individually and they compound into a noticeable trust gap versus competitors who skip them.
What we got wrong on the first launch
Two misses caught in week three. First, the property pages were rendering EPC ratings as text but not as the visual coloured-band imagery that consumers actually recognise from the official certificate format. Search Console clickthrough data showed the listings were ranking but underperforming on CTR relative to similar listings on the national portals. We added the visual EPC band to the page template in week three and CTR closed about 70% of the gap. Second, the valuation form had four fields on the first version and conversion data suggested the third field (estimated rental income) was the drop-off point — landlords often did not have a confident number to put in and abandoned. We moved that field to optional and conversion rate on the form rose 22%.
The PropertyData reverse-sync
A subtle technical choice that paid back: when a property let through the new website, our PropertyData webhook triggered a status update back to the CRM, which then propagated to the other listing channels (Rightmove, Zoopla) Adesh was syndicating to. Pre-launch he was manually marking properties as let in three different systems, with predictable lag and the occasional embarrassing inbound enquiry about a property that had been let two weeks earlier. Post-launch the reverse-sync handles the propagation automatically and the let-status data is consistent across every channel within minutes. The integration code is about 40 lines and has not needed maintenance.
The schema lift specifically
A noteworthy detail for anyone building lettings-agent or estate-agent sites. The previous agency-template platform emitted a single Organization schema entry for the agency. The new site emits 28 RealEstateListing entities (one per active property) with per-property addresses, prices, floor areas, EPC ratings, council tax bands, available-from dates, deposit amounts, and three images each. Google's rich-results for property queries pick up the structured-data entities individually; a property indexed with full RealEstateListing schema is materially more likely to appear in the local-pack rich variant than the same property without the schema. We attribute roughly 40% of the ranking lift to the schema enhancement specifically.
If you have a similar business
If you are an independent lettings or estate agent on a generic agency-template platform that is failing to index in Google, the East London Lets template applies. A Pro-tier multi-page build (£1,499) with a dynamic property route fed from your existing CRM via API, per-property RealEstateListing schema, hyper-local copy naming the streets and developments you specialise in, the regulatory trust signals (ARLA, TPO, CMP, DPS) prominent and schema-tied, and a multi-step valuation form routed back into your CRM via webhook. Pre-condition: your CRM exposes a feed of active listings via API (PropertyData, Reapit, Jupix, Acaboom all do this). The cancellation of the agency-template subscription typically pays for the build inside the first 3-4 months on its own; the inbound enquiry uplift is incremental.
Final note on the hyper-local copy
The hyper-local layer was the cheapest piece of the build and arguably the highest-impact. Each area page (E5, E8, E9) names specific streets, developments and landmarks the agent actually services — Clapton Square, Hackney Downs Park, Hackney Central station, Mare Street, Lower Clapton Road. The copy itself was drafted in roughly four hours from a 20-minute walk Adesh did with the writer through the neighbourhoods. Google's local-search algorithm reads the named entities and uses them as soft signals for geographic relevance; for hyper-local searches like "letting agent E5 Lower Clapton" the specificity wins ranking against national portals that cannot replicate it without sounding generic. Local copy is competitive moat on a small budget.