Discovery findings for the BuzzWatch SMB redesign
Block 0 of the BuzzWatch SMB redesign: a competitive benchmark, what Alberto's funnel data asks for, and a measurement framework, the evidence the homepage and dashboard redesign is built on.
What we found
- 1The leak sits after the free report, not at checkout. Owners come back (about 24% the next day) but don't act and don't pay. So the free week has to be designed as the paywall: a week of actions, then the next ones paid.
- 2Roughly 8 in 10 people who start the report flow leave before their report, and where they leave isn't tracked yet.
- 3One tool already speaks to owners the way BuzzWatch wants to: gotheref, with plain language, simple actions and an instant chat. It's the closest competitor, for now only in French. It confirms the direction, and a few of its mechanics are worth borrowing.
Price in context
Monthly price of the tier each tool presents as right for a small business, as each source lists it. Currencies are not converted.
BuzzWatch
free first week, not final
gotheref
plus a 49€ one-time unlock
Peec AI
entry tier
Otterly.AI
Standard, labeled “for SMEs”. Lite at €29 exists, mostly monitoring
AthenaHQ
labeled “for Self-Guided SMBs”
An agency
the owner's real alternative
Competitive Benchmark
Five AI-visibility tools, tested hands-on as a small business owner would use them. None speaks to an owner at an owner's price.
In the owner's mind the alternative is an agency at €300 to €400 a month, not these tools. So this benchmark is about product mechanics: what to borrow, what to skip.
Positioning
Comparison
- Worth borrowing
- Mixed
- Friction for an owner
AthenaHQ
- Onboarding
- Friction for an owner: Audit sent by email, not live
- Who it speaks to
- Friction for an owner: Marketing and SEO teams
- Pricing
- Mixed: Open pricing, from $295
- First value
- Friction for an owner: Slow to fill, unexplained credits
- Talking to the team
- Sales only
Profound
- Onboarding
- Friction for an owner: Sales questions before any value
- Who it speaks to
- Friction for an owner: Enterprise, B2B jargon
- Pricing
- Friction for an owner: No self-serve paid tier
- First value
- Friction for an owner: Build an agent, not see a score
- Talking to the team
- Sales only(demo)
Peec AI
- Onboarding
- Friction for an owner: Card required before real results
- Who it speaks to
- Mixed: Marketing teams, clean tone
- Pricing
- Mixed: €85, 7-day trial, add-ons
- First value
- Worth borrowing: Tour shows a real AI answer
- Talking to the team
- Sales only
Otterly.AI
- Onboarding
- Worth borrowing: 4 short steps, AI drafts the setup
- Who it speaks to
- Friction for an owner: Marketing pros, analyst jargon
- Pricing
- Mixed: €189 “for SMEs”, add-ons
- First value
- Mixed: Demo report while yours builds
- Talking to the team
- Instant chat
gotheref
- Onboarding
- Worth borrowing: 2 questions that change the flow
- Who it speaks to
- Worth borrowing: Plain French, for owners
- Pricing
- Mixed: 79€, a 49€ unlock competes
- First value
- Worth borrowing: 1 engine free, hidden fixes counted
- Talking to the team
- Instant chat
Evidence per competitor
An “AEO/GEO command center” for marketing teams, from $295/month for its “self-guided SMB” tier. Its content tools are the closest thing to BuzzWatch's action plan, but worded for a content team.
The most enterprise of the five: sales questions before any value, and a product built around an AI agent rather than a report. Its narrow free trial is framed as a step before a demo.
Closest in tone of the four named competitors, but the strictest gate: a card before any real result. Its best idea: every score traces back to real, dated AI answers.
A short setup with no payment step, then a dense, analyst-grade report for a marketing team. A self-described SME tier at €189/month, with per-engine add-ons and five usage meters.
Not on BuzzWatch's list, but the closest product to the same owner: plain language and the most honest paywall found. Not cheap (79€/month), and its paid plan shows all 29 actions at once.
Verdict
What to adopt
- Simple actions an owner can do alone, in a few minutes, with no technical skill. Most are AI-assisted through two simple buttons, not a conversation with a bot (gotheref).
- A free tier with one complete, believable result, not a locked preview (gotheref measures one engine, honestly labeled).
- AI drafts the searches, the owner corrects them, labeled as a draft right where they appear (Otterly).
- One onboarding question: in person or online? It changes what follows. Physical businesses get a Google Maps section with its own actions: fix the listing, answer reviews (gotheref).
- Show the real questions being tested while the report loads, instead of a spinner.
- A plain-language “what's holding you back” first, which BuzzWatch's dashboard already has. Keep it.
- “What to copy, what to fix”, the framing that already works with owners: what the competitor cited instead does, and what's blocking you.
- Technical fixes without a developer, through a direct site connection, with the owner approving each change. A later direction, in line with BuzzWatch's WordPress publishing roadmap (seen at gotheref).
What to avoid
- Sales questions before any value (company size, “are you an agency”).
- A profile question asked just to ask it, when it changes nothing next.
- Unexplained quotas (credits, per-engine add-ons, API meters) on top of a plan. BuzzWatch's searches become one if they aren't explained.
- A card before any uncapped result.
- Complicated interfaces, dense dashboards built for a marketer: share of voice, sentiment, visibility indexes (Otterly, AthenaHQ, Profound).
- Every possible action at once instead of a curated set.
Where BuzzWatch can differentiate
- Price against the agency, not these tools: ~€50 against €300 to €400 a month.
- A curated action plan, one week at a time, instead of a long list of actions all at once. Less to take in for an owner, and it fits the offer: the first week's plan is free, the next ones come with the paid plan.
- Plain language, in the owner's own language. gotheref speaks plainly, but only in French.
- A person one tap away where owners get stuck, by chat or WhatsApp, without going through sales. It's also how BuzzWatch gets the feedback it's missing most.
Where It Leaks, What We Change
The leak sits between the free report and the plan: owners come back, but don't act and don't pay.
leave the report flow before their report
During a ~4-minute flow. Where they leave isn't tracked yet.
of free report viewers come back the next day
None of them had a subscription.
of free accounts reach a paid plan
Checkout itself converts at ~25%, so the leak sits before it.
From Alberto's funnel data: Mixpanel and Stripe, 21/06 to 18/09/2026, on the previous entry plan without a free week. The report flow is today's /discover tool.
The free week is the paywall: one week of actions free, the next ones paid
Day 0
The owner sees their score and the free report. The free week starts.
Design: Hand over this week's actions, curated and doable now.
Email: The week's actions and the date of the next analysis.
Days 1 to 6
Most returns happen on days 1 to 3, on the same numbers: visibility is measured weekly.
Design: Say when the next measurement lands, and show the actions left to do.
Email: Mid-week: the actions left, one tap to mark one done.
Day 7
A new analysis runs automatically. The owner opens it and sees what moved.
Design: Show what their actions changed, then next week's actions, visible but locked behind the plan. The conversion moment.
Email: Your new analysis is ready: what moved, and what comes next.
Every week after
On the paid plan, a new measurement and a new curated set of actions each week.
Design: Prove value through actions done and what they moved.
Email: A weekly recap and the new actions.
Risk to design for: site fixes can be checked right away, but AI answers can take weeks to change. The day 7 analysis has to show progress even if AI citations haven't moved yet.
Searches run the whole model, and they're the hardest part
More searches tracked
the questions customers ask AI
More visibility, more reliable numbers
small samples jump week to week
A higher plan
~€50 to start, then priced by searches
If searches are the unit of price, they have to be understood. Otherwise BuzzWatch falls into the trap the benchmark lists under “avoid”: an unexplained quota, like its competitors.
Proposal
Never ask an owner to size anything
The entry plan comes with a base number of searches, already filled: AI drafts them from the free report, and the owner only corrects them, as questions their customers ask, not as a number. Adding searches is then something the product recommends when there's a reason, in plain words: “Your competitors are cited on 12 questions you don't track yet,” or “With 10 questions your score moves by chance, 20 would make it reliable.” Owners add searches because of a gap they can see, not because of a counter.
Design direction
- AI drafts the searches from the site, the owner corrects them, labeled as a draft where they appear (Otterly).
- A few plain questions shape them: who the customers are, what sells most (Peec AI's audience step).
- Explain the mix in plain words: a few broad questions, mostly niche ones, the way customers really ask (20–30% broad, 60–70% niche).
- Show the gap in questions, never a bare counter: “You're cited in 8 of the 24 questions your customers ask. Here are the 16 where competitors are cited instead.”
- Start from what the dashboard already has: the tracked buying questions, cited or not, and how many more could be tracked.
Where to ask why
Feedback is the priority today because the funnel leaks. Each leak gets a way to talk to the team and a one-tap “why”. If conversion clearly improves, these prompts can step back.
- 1Leaves the report flow before the report“Stuck? Message us on WhatsApp” and a one-tap reason.
- 2Sees the report, no account“A question about your report? Talk to a person.”
- 3Comes back to a report that hasn't moved“What were you hoping to see today?” and WhatsApp.
- 4Stops at the plan choice or checkout“Not sure what the plan tracks for you? We'll explain in 2 minutes.”
- 5Skips an actionA one-tap reason: too technical, no time, not relevant.
Measurement Framework
Each design block ships with the number it should move: one main metric, and one early signal to read during a campaign wave.
Early signal: a step earlier in the same journey. It moves within days and with more people than the main metric, so a campaign wave can be read before the paid numbers come in.
Dashboard
Day 7New model
Search setup
Contact points
Report flow
Homepage
To define before setting targets
- Activation
- Proposed: at least one action from the weekly plan done in the first week, since the numbers only move once the owner acts.
- Retention
- Window and definition to agree: still on a plan after 30, 60 or 90 days, or still doing weekly actions.
Your Brief, Answered
The seven problems from section 3 of your SME brief, “Where they get lost”, each with a design answer.
- 3.1“Why is my report not updated?”
Owners come back to the free report expecting it to update. They don't see that a plan keeps the tracking going.
What we propose: On every return, a status line before the score: when it was measured, when the next measurement lands and why it takes time. Then a few actions to do meanwhile.
- 3.2Searches, slots and quota are confusing
A paying customer didn't know whether to swap her 28 prompts or let them run. Only part of the searches run each week, and none of it is visible.
What we propose: AI drafts the searches from the site and the owner only corrects them. More are recommended from a gap they can see, never from a counter.
- 3.3They want proof before paying
Owners hesitate to pay before seeing what they get. The free report has to show the value and what the plan adds.
What we propose: A complete result on a small scope. Every locked block says what it holds, how much, and which offer opens it.
- 3.4Technical fixes feel out of reach
A slow or badly built site leaves them asking “so what do I do?”. A developer costs more than they can spend.
What we propose: Fixes written out, ready to paste, and sent to the developer in one tap. Later, a direct site connection that applies them.
- 3.5Pauses and churn
Owners stop using the product before they cancel, often after a blocker they can't fix, like a PageSpeed score of 20/100.
What we propose: Acting gets a visible reward before the next measurement: points to win per category, and zero-skill actions to tick off. Points have to be credible.
- 3.6“I have the data, but what do I do now?”
Non-technical owners see their visibility but don't know how to act on it. The team sometimes adds their searches by hand.
What we propose: Open on what's holding them back, in one paragraph. Then fixes ranked by what they earn, each one already drafted so the owner only approves it.
- 3.7Trust and ownership
Some owners were locked out of their own accounts by an agency, and distrust anyone who touches them.
What we propose: “You keep ownership of your accounts and data” wherever we ask to connect something, and the owner approves every change before it goes live.
After Today
To validate today
- 1The direction: the six design blocks below, in this order.
- 2The free week starts automatically at signup, no card. Agreed?
- 3Activation = at least one action from the weekly plan done in the first week. Agreed?
- 4AI fills the entry plan's searches, and more are only recommended when a visible gap justifies it?
Then, one loop per wave
- TristanWireframes. The blocks below, to agree on structure before visuals.
- TristanHTML mockups. The redesigned report flow, dashboard and homepage.
- AlbertoBuild. Implemented with its tracking events, one per step of the report flow included.
- Campaign waveTest. Real traffic on the redesign, read by source.
- TogetherDecide. Read each block against its number, keep what moves it.
The design blocks
- 1
The dashboard, with the action plan first.
The screen owners live on, during the free week and after. This week's actions first, then everything today's dashboard shows, made simpler: what's holding you back, plan and key numbers, tracked buying questions, AI visibility with a competitor ranking, site health and traffic. Plus when the next measurement lands.
Target: Activation: 1+ action done in week 1
- 2
Day 7, the conversion moment.
The new analysis shows what their actions moved, then next week's actions, visible but locked behind the plan. The free week's emails lead here.
Target: Free week → paid plan
- 3
A simple search setup.
AI drafts the searches, the owner corrects them, and the link between searches, reliable numbers and price is explained in plain words.
Target: Setup completed without help from the team
- 4
A contact point on every leak.
Chat or WhatsApp and a one-tap “why”, designed into each step where owners can get stuck.
Target: Conversations started, and conversion of those who talked
- 5
A report flow measured step by step.
The first wave shows where the 8 in 10 leave. The account step or the loading screen is then redesigned first.
Target: Starts the report flow → sees the report
- 6
The homepage.
The outcome in plain language, and the contrast with an agency's price.
Target: Visitor → starts the report flow
Open Questions
Raised during discovery, to bring to Alberto and Guido.
Access and data
- Which variant of the running A/B tests (landing hero v3 vs. v4, leaderboard vs. transcript report format) should be treated as the baseline for the redesign to build on.
- Whether an email or account is asked before the free report today, and why (cost per analysis, abuse protection, follow-up), since it may account for part of that drop.
- A way to track returning free report viewers who reach checkout, the fastest signal that the first-week redesign works.
- How long does a change to a site take to show in each AI engine? A real delay can then be written into the product instead of an estimate.
Raised by the competitive benchmark
- gotheref shows an estimated monthly-visits gain per fix ("+70 visites/mois," broken down per channel) directly on its content page. How would BuzzWatch calculate an equivalent number credibly, given the small-sample-size honesty already agreed for visibility itself?
- gotheref shows a Google Maps ranking for specific local search terms. How is that computed, and is an equivalent achievable with BuzzWatch's own data and access? Most relevant for the local-services segment, which cares most about "best X near me" visibility.
Scope and pricing
- The entry plan is set around €50/month with a free first week, then priced up with the number of tracked searches, not yet final. The exact number and the shape of the free week (fully free or a symbolic price) still need to be settled.
- Every free signup gets a week of actions, and technical suggestions are reviewed by hand today. Can that review follow the volume, or should week 1 hold only simple, non-technical actions?
- Does the refresh have to wait a full week? A new measurement after 3 or 4 days would bring the proof closer to when owners come back. It's a cost question: how much does each extra run cost per free user?