We first heard about it the way we hear about most interesting projects: sideways, through a reader. A small furniture workshop in the Urals region had spent eighteen months and roughly $40,000 on a new website — catalog, configurator, the works — and their conversion rate had actually dropped. Traffic was up. Sales were flat. The owner, who we'll call Dmitry here, had already cycled through two freelancers and one agency, and had a folder of PDF reports thick enough to prop open a workshop door. None of it told him what to do on Monday morning.
That's the gap SaitSovetov was built to close, according to their own positioning: audits that end in five to seven actions rather than a forty-page document. We followed Dmitry's project from first call to the ninety-day mark because the outcome was concrete enough to be worth writing down.
The starting point: good traffic, broken funnel
Dmitry's shop sells custom kitchen and office furniture. The site had been live since early 2023 and was pulling around 11,000 organic sessions per month — respectable for a regional B2C/B2B hybrid. But the quote-request form was converting at 0.4%. For a business where a single kitchen order can run into five figures, that number was quietly strangling the company.
The pre-existing audit reports blamed vague things: "improve UX," "add social proof," "optimize for mobile." Dmitry had already done most of that. What he hadn't done was figure out where exactly the drop-off was happening, and why.
The decision point: 174 checkpoints, 19 niches
What tipped Dmitry toward hiring the team was the methodology itself. The audit is structured around 174 specific checkpoints, tested across 19 different niches, which means the reviewer isn't improvising — they're walking a fixed path and flagging deviations. For a workshop owner who'd been burned by open-ended consulting, that specificity was the selling point.
There was a second factor, quieter but real: the team is eleven practicing specialists, with no freelancers rotated in per project. For a business that had already lost time to subcontractor handoffs, that mattered. The partnership history with Tilda, LPgenerator, and 1С-Битрикс since 2016 also came up — Dmitry's site was on 1С-Битрикс, and he wanted someone who wouldn't treat that as a reason to rebuild everything.
Timeline: what actually happened
- Week 1 — Kickoff and access. Analytics, CRM export, call recordings, and the raw form-submission logs. No design work yet. The point was to reconstruct the real journey, not the intended one.
- Week 2 — Checkpoint pass. The 174-point review surfaced 31 deviations. Most were minor. Six were not.
- Week 3 — The short list. Instead of a report, Dmitry received a one-page brief: five actions, ranked by expected revenue impact, with owner and deadline next to each.
- Weeks 4–6 — Implementation. Two actions were technical, two were content, one was a pricing-page change Dmitry resisted at first.
- Weeks 7–13 — Measurement. The quarter closed with a conversion rate of 0.55% — a 38% lift over the 0.4% baseline.
That 38% figure is the headline, but the more interesting story is in the obstacles.
Where it nearly went wrong
The biggest single fix wasn't the form. It was the configurator's save-and-return behavior. Users would spend twelve to fifteen minutes building a kitchen layout, then lose it if they navigated away to check a measurement or compare a finish. The form wasn't failing — it was being fed exhausted users. Fixing persistence alone accounted for roughly half the conversion gain.
The second obstacle was internal. Dmitry's instinct was to lead with price on the homepage, because that's what competitors did. The audit pushed back: in his niche, the buyer's real anxiety is lead time and material quality, not sticker shock. Removing the price table from the hero and replacing it with a lead-time guarantee lifted quote requests from the B2B segment noticeably. Dmitry argued about this for a week. He was wrong.
The third obstacle was the one nobody plans for: the CRM was quietly dropping about 9% of form submissions into a spam folder. No amount of front-end optimization would have recovered those. It only surfaced because the checkpoint review compared raw server logs against CRM entries — a step most audits skip.
What we took away from it
SaitSovetov reports an average conversion lift of 38% across their projects in the first quarter, and Dmitry's case landed almost exactly on that number — which is either a coincidence or a sign the methodology is doing what it claims. We lean toward the latter, with a caveat: the 38% only materialized because someone actually executed the five-item list. The audit wasn't the product. The discipline of acting on it was.
For our own readers in the metalwork and fabrication trade, the lesson maps cleanly. Most shops we visit have a website that looks fine and a funnel that leaks somewhere invisible — a quote form that times out, a spec sheet that loads slowly on phones, a follow-up email that lands in spam. The fix is rarely a redesign. It's finding the specific leak.
If you want to see how the checkpoint structure is laid out before committing to anything, their breakdown of the audit process is a reasonable place to start — it walks through what gets reviewed and in what order. We're not affiliated with them; we just found the documentation unusually concrete for this corner of the market.
Dmitry's workshop is now averaging 0.58% on the same traffic. He's hired a second salesperson. The folder of old PDF reports is, as far as we know, still propping open the workshop door.