BumbleB Crunch™ in action
Real questions. Real investigations.
Every clip is one question asked in plain English and the investigation BumbleB Crunch runs to answer it — no edits, no narration. Each one ends with how BumbleB Crunch read the question and how it counted.
Start here 5
Websites and ad spend
Ad spend next to on-site behavior: which campaigns bring real sign-ups, where the funnel really leaks, which source is worth the money.
The spend that comes back
“We spent the same on our Google and Meta campaigns. How many sign-ups did each bring in, what was our cost per sign-up on each — and is the cheaper one the one that looked cheaper per click?”
Ad spend joined to website behavior: $12,000 each, but Google brought 960 sign-ups to Meta's 480 — $12.50 a sign-up against $25.00. Meta looked cheaper per click ($0.75 vs $2.00); the cheap clicks were the expensive sign-ups.
The leak is the form
“Our landing page isn't converting. Are we losing more visits at the button, or at the form?”
The leak is the form, not the button: 6,000 visits never click, but 11,200 click and then don't submit. 2,800 make it all the way through.
The source worth the spend
“Which traffic source's visits turn into sign-ups most often — and is it the one sending the most visits?”
Volume and quality name different sources: paid sends 16,000 visits to organic's 8,000, but organic converts at 30% against paid's 7.5%.
The loss after the read
“Of the visits that read a page all the way down, how many get in touch — and where do we lose the most?”
The loss is after the read, not before it: 4,750 of 16,000 deep readers submit the form, while 11,250 read to the bottom and leave. Only 4,000 drop before reading.
The page that holds attention
“Which of our pages actually hold attention, and which just get a glance? Rank them by read-through.”
Pricing holds attention best — 70% of its visits read to the bottom — against 45% for the neighborhood guide and 20% for the blog post. Every page has enough sessions to rank.
Shopify & e-commerce 2
Investigations on a store
The store’s pixel joined with orders, inventory and the rest of a merchant’s data — why sales moved, not just that they did.
Out of stock, not out of favor
“Wine revenue fell from September 10 to 24 compared with August 18 to September 1. By how much, and which wine carried it?”
Two turns across five datasets — orders, inventory, products, refunds and the store's pixel: wine revenue fell from $10,664.30 to $5,774.20, almost all of it one rosé. Its views held (4,426 vs 4,423) but stock sat at zero all 15 days — the data supports out of stock, not lost interest.
The checkout that loses phones
“Walk me through our store funnel — product views to cart to checkout to purchase. Where do we lose the most people?”
7,000 view a product, 5,000 add to cart, 3,600 start checkout — and only 1,320 buy. The biggest loss is the last step, 2,280 people. Split by device: 60% of desktop shoppers who start checkout finish it, against 25% on mobile.
Real estate & marketplaces 2
Investigations on a property marketplace
Clickstream joined with enquiries from the apps and the listings — where leads went, and why.
The enquiries that vanished
“Rental enquiries dipped — split by city, then Pune by app, then find when the Android ones stopped. Did Android users lose interest, or did the enquiry step stop working?”
Six short questions across the clickstream, the app's enquiry records and the listings: Pune's Android rental enquiries went 42 → 28 → 0 by March, the last on Feb 16 — while views held at about 85 a day. The pattern fits a broken enquiry step, not lost interest. Speeded up — the full conversation took about 45 minutes.
Not the ads — the biggest client's listings
“Leads fell in March. Was it our ads — or something else? Seven small questions, from leads per day to the client that moved.”
Six datasets — clickstream, app enquiries, listings, clients, Google and Meta exports: leads fell about 20% a day (124.6 → 99.6) while ad spend and paid visits held flat. One client fell from 38.3 to 12.5 leads a day after taking 40 of its 120 listings down on March 10; without it, leads were roughly flat. Time-lapsed — about 75 minutes of analysis.
Behavior, natively 4
Investigations on product data
A hidden leak under a flat line, a drop with a name, a north-star metric carried by a few heavy users.
The founder's conversation
“Growth feels flat and I can't tell where we're losing people — I don't even know the right question to ask. Where should I focus?”
Three turns: BumbleB Crunch turns a vague worry into first questions, lays out a plan before running anything — flagging that traffic-source data is empty — then runs the funnel and names the leak: 12,300 lost between viewing a listing and contacting an agent.
The leak under the flat line
“Walk me through our funnel, from searches to viewing a listing to contacting an agent. Which step loses the most people?”
Of 20,000 searchers, 16,000 view a listing but only 3,700 contact an agent. The leak isn't at the top — 12,300 are lost between the listing and the contact.
The north star, decomposed
“The median and 90th percentile of listing views per person, and the ratio between them — are a typical user and a heavy user close or far apart?”
Far apart: the median person views 4 listings, the 90th percentile views 19 — a 4.75× gap. A thin power core carries the metric.
The loudest channel isn't the best
“Which traffic source's visits turn into agent contacts most often — and is it the same source that sends us the most visits?”
Not the same source. Referral converts best at 36% (paid 18%, organic 9%), while organic sends the most visits — 19,000 of 42,500. Organic was just loud.
Any dataset 2
Beyond clickstream
Your own databases and combined datasets — orders, listings, ads — investigated with the same reasoning.
The flat quarter, split
“Revenue's flat quarter over quarter. Overall, what changed underneath — the number of orders or their size — and in which region is the order size moving?”
An orders table, no tracking involved: orders rose from 8,000 to 8,800 while the average order shrank — and it's all the West, down $40 an order while every other region rose.
When insiders buy together
“Counting open-market purchases from January through May 2026 by directors and officers: which five company-months had the most different insiders buying — and for those companies, how many bought, in how many months, on how many days?”
Real SEC Form 4 filings — 108,302 filings, three tables joined. The Plan agent splits it into four asks; the Analyst runs them: Taiwan Semiconductor leads with 28 insiders buying in both April and May, then F&M Bank, Everforth and Vital Farms. Speeded up — the full run took about 14 minutes.
Ask your own question next
Point BumbleB Crunch at your own data and watch it investigate — free to start.
Sign up free