The agency said six weeks. It took four months. This is not a story about a bad agency or a difficult client -- it is a composite drawn from multiple real NZ accounting firm redesign projects, and the patterns repeat with uncomfortable consistency. The brief was detailed about the wrong things, the content arrived late, and the old site had been quietly failing in ways nobody had noticed.
The brief ran to fourteen pages. It specified the exact shade of blue for the header (Pantone 2955 C), the preferred typeface (Montserrat, because a partner had seen it on a site they liked), and the requirement for a "modern yet professional feel." It did not specify what the firm wanted the website to do.
This is the standard pattern. Accounting firm website briefs focus heavily on aesthetics and say almost nothing about function. Who visits the site? What do they need when they arrive? What should happen after they leave? These questions went unasked and unanswered until month two, when the agency raised them and the project stalled while the partners debated.
The brief also contained a detailed sitemap -- twenty-three pages mapped out in a neat hierarchy. It had been assembled by listing every page on the existing site and adding a few new ones. Nobody had assessed whether the existing pages served any purpose. Nobody had checked the analytics to see which pages actually received traffic. The sitemap was a mirror of the old site with a fresh coat of paint specified.
Accounting practices run on precision. Tax returns are exact. Trust accounts reconcile to the cent. Somehow, this rigour evaporates when the deliverable is a website. The brief was detailed about things that do not matter and silent about things that do.
The firm approached three agencies. The quotes came back at $8,500, $16,000, and $27,000. All three proposed a WordPress build with broadly similar page counts. The $27,000 proposal included a content strategy workshop and brand guidelines document. The $8,500 proposal was three pages long and listed deliverables without explanation.
The firm chose the middle quote, reasoning that the cheapest was too cheap and the most expensive was excessive. This is how most professional services firms select web agencies -- not through rigorous evaluation, but through a pricing heuristic that assumes the middle represents sensible value.
What the firm did not do was compare portfolios in detail, check references from comparable clients, or ask each agency to explain their process for handling content delays. These questions would have revealed that the selected agency had never built a site for an accounting practice and had no understanding of the compliance obligations that come with marketing professional services.
The agency was competent. They had built attractive websites for hospitality, retail, and technology clients. But an accounting firm's website has specific requirements -- Chartered Accountants Australia and New Zealand guidelines on advertising, tax agent obligations, and the professional tone that clients of accounting firms expect. None of this appeared in the agency's proposal because they did not know to include it.
The agency's process began with a "discovery workshop" -- a two-hour session where they asked the firm about their target market, competitive advantages, and brand values. The partners attended and gave thoughtful answers. The agency produced a twelve-page discovery document summarising what they had learned.
Every insight in that document was information the firm already knew and had, in fact, included in their original brief. The discovery phase discovered nothing. It reformatted existing knowledge into the agency's preferred template. This cost the firm approximately $2,000 in billable hours and two weeks of elapsed time.
Discovery phases have genuine value when they uncover things the client does not know -- user behaviour data, competitive positioning gaps, technical requirements the client had not considered. In this case, the agency asked what the firm wanted rather than investigating what the firm's clients needed. No user research. No analytics review. No competitor analysis beyond a superficial glance at three local firms' websites.
The firm accepted the discovery document without comment, assuming the agency knew what they were doing. This assumption persisted for approximately six more weeks before it became apparent that the agency was building based on the firm's stated preferences rather than any independent assessment of what would actually work.
The agency presented three homepage concepts in week three. The designs were polished and professional. The partners reviewed them in a meeting and selected Option B -- a clean layout with a prominent services grid and a muted colour palette. The decision felt conclusive.
Two days later, the managing partner viewed the selected design on their phone during a commute. The services grid, which looked elegant on a 27-inch monitor, collapsed into a single column on mobile. The text was small. The call-to-action button required scrolling. The partner emailed the agency requesting "significant changes to the mobile experience."
This is the point where timelines begin to diverge from plans. Mobile design is not a secondary consideration -- over 55% of NZ web traffic comes from mobile devices according to Stats NZ digital access data. But most design presentations happen on desktop screens in meeting rooms, creating a systematic bias toward layouts that work on large displays.
The agency spent a week revising the design. The revised version addressed the mobile concerns but altered the desktop layout enough that another partner, who had preferred Option A, reopened the design discussion. Two more rounds of revision followed. The design phase, scheduled for two weeks, consumed five. Each revision was reasonable in isolation. Collectively, they added three weeks and approximately $3,000 in additional scope.
The agency requested all page content by the end of week three. The firm's office manager was assigned the task alongside their regular duties. She requested copy from each partner for their respective service area pages. By week three, she had received one complete page (tax advisory, from the most junior partner) and a bullet-point list for another.
Content delays are the single most predictable failure mode in professional services website projects. Partners are busy with client work. Writing web copy feels less urgent than everything else on their desk. The deadline passes. Then another deadline passes. The agency sends polite follow-up emails that get filed for later.
By week six -- when the original timeline had the site launching -- the firm had supplied content for four of twenty-three pages. The agency had begun development on the pages they had content for, leaving the rest as wireframes with placeholder text. The project entered a holding pattern.
The content eventually arrived in week nine, delivered in a series of Word documents with inconsistent formatting, varying tones of voice, and several sections that contradicted each other. Two partners described the same service in fundamentally different terms. The agency's copywriter spent a week reconciling and editing the content. This editing cost appeared as an additional line item that surprised no one at the agency and everyone at the firm.
With content still arriving in fragments, the agency made a decision that project managers in web agencies make routinely: they started building. The homepage and services pages were developed against the approved design while inner pages remained in wireframe. The logic was sound -- keep the project moving, use the available content, and slot in the rest as it arrives.
The problem emerged when the design revisions from the mobile feedback cascaded into pages that had already been built. The revised header height affected the homepage hero. The new navigation structure required restructuring the services dropdown. Two weeks of development work needed partial rebuilds that would not have been necessary if the design had been finalised before coding began.
This is a tension that exists in every web project. Waiting for perfect design sign-off delays development. Starting development before sign-off creates rework risk. Experienced agencies manage this tension through modular builds that isolate design-dependent elements. This agency built sequentially, meaning changes propagated through everything.
The rework added approximately $1,500 in development time. The agency absorbed some of this cost and billed some of it. The firm disputed the bill. A tense but professional conversation followed. Both parties were partially right. Neither was entirely satisfied. The project continued with a slightly strained relationship that would persist through to launch.
By the end of month two, the project had accumulated a list of additions that nobody had formally approved or quoted. A staff portal page that was not in the original brief. A newsletter signup form with Mailchimp integration. A careers section with application functionality. A client testimonials carousel. Each addition was small. Together, they represented roughly $4,000 of additional work.
The scope creep conversation is one that agencies dread and clients resent. The agency did not raise it early because each individual request seemed minor and saying no felt like poor client service. The firm did not realise the additions were outside scope because they assumed a $16,000 website would include "basic" features like a careers page.
Both assumptions were wrong. A careers page with application functionality requires a file upload system, an email notification workflow, and compliance with the Privacy Act 2020 for handling applicant data. That is not a basic feature. It is a development project within a development project.
The formal scope change discussion happened in week ten. The agency presented an itemised list of additions totalling $4,200. The firm pushed back on several items they considered standard inclusions. A compromise was reached at $2,800 in additional charges. The goodwill cost was higher. The firm felt nickel-and-dimed. The agency felt unappreciated. Both wrote it off as the cost of doing business.
The old site had been live for eight years. In that time, partners and staff had added pages without any governance process. Blog posts that were never part of a content strategy. Event pages for seminars held in 2019. PDF uploads -- tax guides, checklists, client newsletters -- stored in a media library that had never been audited.
The agency's content migration plan assumed twenty-three pages. The actual site contained over two hundred URLs when the developer ran a crawl. Most were blog posts, many were empty category pages generated by WordPress, and some were draft pages that had never been published but existed in the database.
Deciding what to migrate, what to redirect, and what to delete required input from the firm. The firm did not have time to review two hundred pages. The agency did not have the context to make the decisions independently. A stalemate developed that consumed most of week eleven.
The eventual compromise was pragmatic. Migrate the core pages. Redirect service-related blog posts to the relevant new service page. Delete everything else and accept the minor SEO impact. This was the right decision, arrived at six weeks later than it should have been. Had anyone crawled the old site during the brief-writing phase, the migration scope would have been clear from the start and the quote adjusted accordingly.
The old site used a permalink structure that included dates and category names. The new site used clean slugs. Every existing URL would break on launch unless redirects were mapped and implemented. The agency included "basic redirect setup" in their proposal. They had budgeted for redirecting twenty-three pages. They now needed to handle over two hundred.
URL redirect mapping is tedious, unglamorous work that has outsized consequences when done poorly. Each old URL needs to point to the most relevant new URL. Old blog posts about tax changes need to redirect to the new tax services page. Old event pages can redirect to the homepage. Category archives can be handled with pattern-based rules.
The agency implemented redirects for the core pages and used a catch-all rule to send everything else to the homepage. This is the standard shortcut. It works from a technical perspective -- no pages return 404 errors. But it damages search engine visibility because Google interprets a mass redirect to the homepage as a signal that the original content no longer exists in any meaningful form.
The firm's search rankings dropped measurably in the month following launch. Several pages that had ranked well for local accounting-related queries disappeared from results entirely. Recovery took approximately six months of the new site building its own authority. The cost of proper redirect mapping would have been roughly $800. The cost of lost search visibility was substantially higher.
During the content migration, the developer examined the existing contact form configuration and found something that should have been caught months earlier. The form's notification email was set to deliver submissions to a staff member who had left the firm eight months previously. The staff member's email address had been deactivated when they departed. Every form submission for the past eight months had been sent to an address that no longer existed.
No bounce notifications had been generated because the email was on the firm's own domain -- the mail server accepted delivery to the local address and silently discarded the messages. The form showed a success message to every user who submitted it. From the client's perspective, they had made contact. From the firm's perspective, nothing had happened.
The firm could not determine how many enquiries had been lost. The form plugin had a submission log, but it had been disabled to reduce database size -- ironically, a decision made by the same staff member who had since departed.
This is not an unusual situation. Contact form routing in professional services firms is frequently set up by someone who later leaves. The handover process, if one exists, rarely includes "check which email address the website contact form sends to." The redesign project surfaced this problem. Without the redesign, it might have continued indefinitely.
The old site had Google Analytics installed. Three years of data sat in the account, unexamined. During the discovery phase, nobody thought to look at it. When the developer finally accessed the analytics during migration planning, the data told a story that contradicted several assumptions the firm had been operating under.
The firm believed their homepage was the most important page. Analytics showed the tax services page received more traffic than the homepage, almost entirely from organic search. People were finding the firm by searching for specific tax services, not by navigating to the homepage first. The new site's information architecture, which subordinated service pages to a single dropdown menu, was actively working against the firm's most effective acquisition channel.
The firm believed clients visited the team page to learn about the partners. The data showed the team page had a bounce rate of 73% -- visitors arrived, looked briefly, and left. The page was not converting interest into contact. It was satisfying curiosity without providing a next step.
The firm believed the blog was pointless because they had only published six posts. The data showed those six posts collectively attracted more organic traffic than the homepage. One post about IRD tax residency rules had been steadily attracting visitors for two years.
None of these insights were complicated to obtain. They required logging into Google Analytics and spending thirty minutes reading the data. Nobody had done it. The redesign proceeded without this information shaping the decisions that mattered most.
The site launched on a Friday afternoon in month four. The agency had pushed for a Monday launch, but the managing partner wanted it live before a networking event that weekend. This is a common and consistently poor decision. Friday launches mean that the first 48 hours of a live site -- the period when bugs surface, links break, and forms malfunction -- coincide with the period when the agency's development team is least available.
Within three hours of launch, the firm received an email from a client noting that the "Book a Consultation" button on mobile devices linked to the wrong page. A partner discovered that their bio contained a qualification they had not yet completed -- content from a draft version that had not been updated. The footer displayed the firm's old phone number.
None of these issues were catastrophic. All were fixable within minutes. But each required the developer to log in, make the change, and clear the site cache -- tasks that took until Monday morning because the developer reasonably did not check email on Saturday.
The following week produced a steady stream of minor issues: a broken link in the services section, a PDF download that returned a 404, and a privacy policy that referenced legislation by its pre-2020 name. Each fix was quick. The aggregate disruption to the firm's week was not. Launch is not an event. It is the beginning of a maintenance relationship.
This project ran over time by ten weeks, over budget by $4,800, and delivered a website that the firm described as "fine, but not what we expected." The agency described the project internally as "difficult client, never again." Both assessments contain truth. Neither is the whole story.
The patterns that produced these outcomes are consistent across nearly every accounting firm website redesign. The brief that specifies aesthetics but not objectives. The content that arrives late. The design approval that unravels on mobile. The old site that harbours surprises nobody checked for. The scope that grows through reasonable individual requests. The launch that happens before the site is truly ready.
These are not failures of competence on either side. They are structural features of how professional services firms and web agencies interact. The firm lacks technical knowledge and treats the website as a peripheral project. The agency lacks sector knowledge and builds what it is told rather than what is needed.
The fix is not finding a better agency. It is approaching the project differently. Audit your existing site before writing a brief. Review your analytics. Check your contact forms. Crawl your URLs. Define what the site needs to achieve, not what it should look like. Set content deadlines with consequences. Launch on a Monday. Budget for ongoing maintenance. Treat the website as a business system, not a marketing expense, and the outcome changes materially.
Every accounting firm website redesign follows a version of this arc. The variables change -- the specific delays, the particular surprises, the exact budget overrun -- but the structure is remarkably consistent. Knowing the pattern does not prevent every problem, but it converts surprises into anticipated events. That alone changes the outcome.
