Magic EdTech builds accessible, inclusive learning products for K–12, higher education and corporate L&D. Two decades of delivery credibility, and a website that search engines read as a generic services site. The content was already there. What was missing was the mapping between what buyers typed and what the site said. We rebuilt that layer, cleared the technical drag underneath it, and made the site readable to AI assistants at the same time.
EdTech services & digital learning: K–12, higher ed, corporate L&D
Enterprise services team, global delivery
United States (primary market)
Established enterprise: multi-vertical learning services
Qualified organic traffic and inbound enquiries from learning-product buyers
Magic EdTech sells to some of the most research-heavy buyers in B2B. A district technology director, a university digital learning lead, a corporate L&D head: none of them fill in a form on first visit. They read for weeks, compare vendors against accessibility and interoperability requirements, and build an internal shortlist long before anyone picks up the phone.
The site had the substance those buyers look for. What it didn’t have was any structural relationship between that substance and the language buyers actually search. Core service pages carried no targeted terms. Deep capability content sat in isolation with nothing linking it to the pages that sold the work. Slow templates meant a share of visitors never saw either.
The brief we were given was broad: more traffic, more leads, better rankings. The first thing we did was narrow it.
Accessibility remediation, content engineering, QA and platform work each had a page, but none of them targeted the terms a district or university team types. The pages described the service in Magic EdTech’s internal vocabulary rather than the buyer’s.
Slow-loading templates, an unstructured on-page setup and a flat site architecture. Crawl budget was spent on low-value paths while the pages that mattered sat several clicks deep with no internal authority flowing to them.
Domain Authority had been flat at 36. A portion of the referring-domain profile was low-quality and actively diluting the trust signal that an accessibility and compliance vendor needs more than most.
Stage 1 examines five dimensions before anything is recommended. Here’s what each one revealed.
Magic EdTech had already produced more genuinely useful material on accessible learning design than most of the vendors outranking them. The diagnosis found no content gap worth solving first.
What it found was a translation gap. The site spoke in delivery language: the words used inside a learning-engineering team. Buyers search in problem language: WCAG remediation, LMS content migration, accessible course design, learning content QA. No page sat between those two vocabularies.
So the recommendation wasn’t a content calendar. It was: make the site speed up, make the pages you already have say what buyers search for, and connect them to each other.
Buying committees research for months. BOFU pages aren’t the last step: they’re the comparison step, and they were the weakest pages on the site.
Rewriting pages on slow, badly linked templates would have capped the result. Fixing the floor first made everything after it compound.
The same signals that help Google classify the brand decide whether an AI assistant names it in an answer. One body of work, two channels.
We didn’t ask for a bigger content budget. We asked for six weeks to fix what was already there.
Specific. No “we optimised your content strategy.” Here’s exactly what happened, in order.
Template-level speed work across the main page types, with fixes specced for Magic EdTech’s own development team rather than handed over as a severity list.
Flattened the depth to key commercial pages, removed indexation waste, and defined a hierarchy that reflects the actual service lines.
One primary query set per commercial page, segmented by buyer stage and by segment (K–12, higher ed, corporate L&D) with cannibalisation resolved before any rewriting began.
Service pages rebuilt around the problems buyers name (accessibility remediation, content migration, learning QA) in the language of the RFP, not the delivery deck.
Existing depth grouped into topical clusters, each with a pillar page and mandatory internal links routing authority into the commercial pages beneath it.
Comparison and capability pages reworked for the shortlist moment, with named practitioner expertise, delivery evidence and accessibility credentials made explicit on the page.
Toxic referring domains identified and disavowed, then replaced with relevant placements in education and learning-technology publications.
Schema expanded across service and resource templates, FAQ blocks written against real buyer questions, and entity coverage broadened so the site reads as a specialist rather than a generalist.
GA4 reconfigured so organic sessions can be traced through to enquiry: the reporting line the marketing team needed to defend the channel internally.
Measured at Month 6 against the pre-engagement baseline. No paid traffic changes during the engagement.
YMYL, local, and competing against NHS.uk. Rather than fight for the head terms, we took the specific high-intent searches a family makes days before booking a visit.
400 blog posts, zero bottom-of-funnel content and no leads. With a 90-day window there was no time to publish, so we triaged what already existed instead.
A genuinely differentiated organic range that search couldn’t distinguish from anyone else’s. We fixed the cannibalisation, worked the products that already sold, then moved authority.
Every case on this site, including this one, started with a diagnosis before any work was recommended. That’s Stage 1 of The Clarity Method™. That’s where we start.
Book a free strategy call. No commitment, no pitch deck.