James & Ella: Funnel & PDP CRO Audit
Marcin Mleczko · 2026-06-03 Audit scope: ella.co homepage, Freeze-Dried Raw PDP (grass-fed beef), reviews page, contact path. Read-only public-facing audit, no analytics access.
How to read this
I've grouped findings by where they live in the funnel, and tagged each with priority for your June dev sprint. Quick wins are <1 day fixes. Structural items are bigger, likely Phase 2 after the presell test reads.
What I couldn't audit without analytics or admin access: - Cart drawer behaviour with items added - Checkout flow + payment methods shown - Mobile sticky CTA behaviour during scroll - Page weight + load speed (needs Lighthouse run) - Actual returns / "what if my dog doesn't eat it" policy text
Those should be tested directly in your June sprint. I've flagged the questions to answer for each.
Top 3 fixes (the 80/20)
If your dev agency only has time for three things this month, these.
1. Surface the £40 trial price + social proof above the fold on the PDP
Right now the PDP says "From £40 (for a trial box)" in the price line, but the actual trial cost is hidden behind the calculator. A cold visitor arriving from a presell or ad doesn't see the £40 anchor clearly, they see a calculator form. Add: "14-day trial · £40 delivered · 900+ five-star Trustpilot reviews" as a strapline directly under the H1, above the calculator. This is a 1-day dev task and probably worth a measurable CR bump.
2. Add a money-back / satisfaction guarantee
Currently risk reversal = "cancel anytime, no strings attached", that addresses sub commitment, not the bigger fear: "what if my dog won't eat it?". The single most repeated objection in the audience research is exactly this. Add: "Doesn't eat it? Email us within 14 days for a full refund of the trial." Bake it into the calculator section + checkout. Even if you never need to honour it, the presence reduces friction.
3. Show the average star rating + total review count everywhere
Trustpilot widget exists but isn't surfaced as a number anywhere visible: - Homepage: no rating - PDP above the fold: no rating - Reviews page: no average shown
You have 933 reviews at 4.9 stars, that's a top-tier signal. Put it in three places minimum: homepage hero strapline, PDP under the H1, footer of every page. This is a one-line code change once the widget is properly embedded.
PDP findings (the highest-leverage page)
Above the fold
P0: Trial price clarity - Current: "From £40 (for a trial box)" → user has to use calculator to see real cost - Fix: Anchor the £40 in a confidence band above the calculator. Example: "Every dog's trial is £40. Calculator-portioned to last 14 days. Cancel anytime." This removes the "is it actually £40 or is it more?" suspicion before calculator entry.
P0: Social proof missing above the fold - Current: Trustpilot link present, no star count or review number visible - Fix: "⭐⭐⭐⭐⭐ 4.9 from 933 verified reviews" under the H1. Use the live widget if possible, even a static strapline is fine.
P1: Eight benefit icons are too many - Current: 8 icons (100% raw, human grade, gluten/grain free, all natural, delicious, made by experts, soft crumbly, no prep) - Issue: Cognitive overload. None of them are the actual hook ("vet-approved" / "British made" / "what kibble can't do"). - Fix: Cut to 3 maximum. The three that matter for cold conversion: British-made · Made by canine nutritionists · 900+ 5-star reviews. Move the rest to a "Why James & Ella" expandable below the calculator.
P1: H1 is a category description, not a hook - Current: "Freeze-Dried Raw Dog Food" - Issue: Generic, doesn't differentiate. Most owners arriving cold don't know what freeze-dried raw is, so the H1 is doing zero conversion work. - Fix: Test a benefit-led H1: "Real raw food. No freezer. No prep." or "The 14-day raw food trial James Middleton built for his own dogs." Keep current as H2 / SEO subhead.
Variant + calculator flow
P0: Too many decisions stacked Current selectors before "Add to basket": 1. Flavour (3 options): Country game / Free-run chicken / Grass-fed beef 2. Bag weight (7 options): 1kg through 10kg 3. Subscription type (4 options): Replacement / Trial / One-time / Subscription 4. Dog age (input) 5. Dog weight (input)
That's five decisions on one screen. Cognitive load is the #1 killer of trial conversion in pet food. Solutions:
- Hide the bag-weight dropdown for trial. The trial portion is auto-calculated, manual bag weight selection is irrelevant for a first-time user. Show bag weight only after subscription option.
- Subscription type → simplify to two clear paths. "Start a 14-day trial" vs "Order one-time / skip trial." Right now "Replacement / Trial / One-time / Subscription" is ambiguous, what's "Replacement"? Why would I "One-time" if I haven't trialled?
- Default flavour to most popular (Beef or Chicken). Don't make new visitors choose between 3 unknowns. Show a small "Most popular" badge on one variant.
P1: Calculator doesn't surface ingredient confidence - Once dog data is entered, the calculator shows: cost/day, grams/day, delivery, discount, total. - Missed signal: "For Hattie (6yr, 12kg) we recommend ~80g/day = ~£2.85/day", turn cost into per-day pence figure to anchor against the kibble bag price comparison.
P1: No "what's in the box" before submitting - Add a small panel: "Your trial contains: 14 days of food, a feeding scoop, a printed transition guide, and your dog's first stickers." Tangible items = perceived value spike.
Trust, reviews, risk reversal
P0: Money-back / satisfaction guarantee missing - See Top 3 #2 above. The fix is policy + copy + placement.
P1: Reviews on PDP are 4 cherry-picked - Currently: Michelle (Lola Cockapoo), Grace (Monkey), Dhelia/Molly, Victoria (Paddy) - Issue: Only 4 visible. The user can't see "wait, do dogs LIKE me have this experience?" - Fix: Add the Trustpilot widget embed with live count + scrolling carousel of recent reviews, not 4 static blurbs.
P1: Press logos are decorative, not clickable - Tatler, Telegraph, Pet Gazette, Daily Mail, The Times, Hello! - Each logo should link to the actual article (or a "press" landing page with all six). Right now they're a static trust-badge wallpaper, the user can't verify, so the brain discounts the claim.
P1: Founder quote on PDP could be 3x longer - Current James Middleton quote is functional but short. - Move part of the founder backstory (depression / Ella / memoir) onto the PDP, it's the brand's strongest differentiator and currently lives only on /about-us.
Section order below the fold
Current order: 1. Highlights → 2. Calculator → 3. FAQs → 4. Ingredients × 3 flavours → 5. Nutritionals → 6. Why subscribe → 7. Benefits → 8. Testimonial × 1 → 9. Press logos → 10. Reviews × 4 → 11. Founder → 12. All FAQs link
P1: FAQs are placed too high. Putting "How does a trial work?" before ingredient/proof sections reads like the page is defensive. Move FAQs below the benefits + testimonials.
P1: Ingredient cards × 3 flavours = 36+ images stacked. Most users only care about the flavour they're considering. Progressive disclosure: show only the selected flavour's ingredients, hide the other two behind a "Compare flavours" toggle. Halves the page weight.
P1: Founder story buried at position 11. This is James & Ella's strongest differentiator vs. category leaders. Push to position 4 or 5, right after the benefits section.
Homepage findings
The homepage is doing brand-positioning work but very little conversion work.
P0: No trial offer surfaced on homepage - The £40 14-day trial is the brand's strongest acquisition hook. It doesn't appear anywhere on the homepage. - Fix: Add a hero banner or below-hero strip: "New here? Start the £40 14-day trial. Calculator-portioned to your dog. Cancel anytime."
P0: No pricing visible anywhere on homepage - First-time visitors can't establish a price anchor before committing to click into a product. They bounce. - Fix: "From £40" or "From £2.85/day" should appear in the hero or first product card.
P1: Calculator entry point is invisible - The "How It Works" section describes "Fill in your dog's name, age and weight" but there's no button to actually do that on the homepage. - Fix: Persistent "Build My Dog's Plan" CTA in the nav bar OR a homepage-level dog-profile widget that captures name/breed/age before sending them to a flavour decision.
P1: Hero subhead is doing too much - Current: "Support your dog's health with the power of Freeze-Dried Raw, the innovation of Cold-Pressed, and the balance of Kibble + Raw..." - That's three product lines in one sentence. New visitors get whiplash. - Fix: Simpler, lead with the hero product (Freeze-Dried Raw) and let the others be discovered.
P1: 6+ competing CTAs - Shop the range / Find out more / How it's made / How it works / Our ingredients / Read all / Shop our products - Too many roads. One primary CTA ("Start my dog's trial") + one secondary ("How it works"). The rest can be footer links.
P2: Product range guidance missing - 4 product tiers (Freeze-Dried, Kibble + Raw, Cold-Pressed, Topper) without a "Which is right for my dog?" quiz or comparison. - Fix: Add a one-question quiz "Currently feeding kibble / raw / mix?" that routes to the most logical first product.
Reviews page findings
This page is the biggest social proof asset the brand is under-using.
P0: No average star rating + total count displayed. - 933 reviews × 4.9 stars should be the page hero. Currently invisible.
P1: No filtering by dog breed, age, or outcome. - The strongest review use-case is: "Show me reviews from owners of fussy eaters / dogs with itchy skin / Cockers." - Fix: Three filter tabs minimum: Dog type · Health concern · Product. Tags drawn from existing review tags or manually curated.
P1: No dates visible on reviews. - Undated reviews lose credibility. Add dates (or relative "8 months ago") to each review.
P2: No video reviews. - Video reviews convert at 3-5x text reviews. Worth a small program: free trial bag in exchange for a 60-second video.
P2: No "load more" pagination. - Showing only ~12-15 reviews when 933 exist undersells the volume. Add infinite scroll or "Load more" pagination.
Funnel congruence, what the presell promises that the PDP doesn't reinforce
This is where most clients lose post-click conversion. The v3 presell we just built establishes claims that the PDP should immediately validate. Currently:
| Presell promise | PDP delivery | Status |
|---|---|---|
| "From £40, 14-day trial" | "From £40 (for a trial box)" | ✓ Matches |
| "Calculator-portioned to your dog" | Interactive calculator | ✓ Matches |
| "Cancel anytime, no strings" | "cancel any time, with no strings attached" | ✓ Matches |
| "4.9 stars across 900+ Trustpilot reviews" | Trustpilot link only, no star/count visible | ✗ Congruence break |
| "Made in the UK by James Middleton and his in-house canine nutritionists" | Founder quote present but no nutritionist credentials | ✗ Partial gap |
| "Bowl was empty before I'd put her water down" (10-day result) | No time-to-result claim anywhere on PDP | ✗ Gap |
| "If she doesn't eat it, you've spent less than the kibble bag you'd be replacing" | No money-back signal | ✗ Gap |
Fix: The PDP's above-the-fold band should explicitly mirror these four lines: rating, founder, time-to-result, refund. Five words each, no more. That's the entire trust scaffolding for cold paid traffic.
Mobile-specific (needs a live device check)
I couldn't fully audit mobile from web fetches. Three things to verify on a real iPhone in your June session:
- Sticky cart / sticky CTA behaviour, does the "Add to basket" stay visible while user scrolls through ingredients? If not, a sticky bottom bar is a P0 mobile fix.
- Calculator interaction on small screen, does the multi-input calculator fit one viewport, or does the user have to scroll while filling? Touch-target sizing on the dropdowns.
- Image carousel performance, 6 hero slides + 36+ ingredient cards = potential mobile speed issue. Lighthouse score will tell you in one click. If LCP > 2.5s, lazy-load the ingredient sections.
Quick wins for the June sprint
If your dev agency has 5 days, here's the punch list in priority order:
- Above-the-fold trust band on PDP, H1 + star rating + 14-day trial price + refund signal. One day.
- Cut benefit icons to 3, half day.
- Add money-back guarantee policy + copy, depends on legal/ops sign-off, but the dev change is half a day.
- Surface Trustpilot widget with live count on homepage + PDP, half day.
- Default flavour selection + hide bag-weight on trial flow, one day.
- Make press logos clickable, half day.
- Move founder story up the PDP section order, half day.
- Add date stamps to reviews on /our-reviews, half day.
Total: ~5 dev days. Should be doable inside your June retainer hours.
Structural items (Phase 2 after presell test reads)
Hold these until you have post-test data. They're bigger lifts and would dilute the test if changed simultaneously:
- Homepage dog-profile capture, quiz-style flow before product selection
- PDP variant flow redesign, collapse five decisions to two
- Reviews page filtering, by dog type, outcome, product
- Video reviews program, operational lift, content lift
- Progressive disclosure of ingredient cards, show selected flavour only
- Cart abandonment flow, needs email + SMS infrastructure
What I didn't audit (open items for June)
Pass these to your dev agency:
- Cart drawer + checkout flow, needs items added + walked through. Specifically: free shipping threshold, payment methods (Klarna/ClearPay), guest checkout availability, delivery date prominence.
- Email post-purchase sequence, does it set 7-day expectation, send transition guide, surface support?
- Subscription management UX, how easy is it really to pause/skip/cancel?
- 404 audit,
/help/returns 404,/contactreturns 404 (works as/contact-us/). Audit broken internal links. - Page speed (Lighthouse), needs a single run to identify image weight + JS bloat.
- Mobile-specific testing as flagged above.
One philosophical note
The brand has an unusually strong asset stack: James Middleton's memoir, the depression/Ella backstory, in-house canine nutritionists, 933 4.9-star reviews, six tier-1 press features. None of that work shows up above the fold on the PDP. A cold visitor arriving from a Meta ad sees a generic category H1 and a calculator form. Everything you'd want to lead with is buried below the fold.
The single biggest CRO move is not a structural redesign, it's pushing your existing trust assets above the fold. That's a one-week dev sprint, not a quarter-long redesign.
Audit prepared for Erica Hyde + James Middleton. Send questions back any time.