RPS
CRO Testing Roadmap
September 2026
Six concepts drawn from the customer segmentation survey of 101 customers, 628 orders, 161 post-purchase answers, 703 published product reviews, and every test we have run so far.
1
Why we're testing
Right now the specs are in the product title, pipe-delimited, six at a time. To compare two pumps a customer has to read two long strings and hold the differences in their head. The product page does have a spec block, but it sits several folds down, and it lists construction features shared by every RPS pump, hermetically sealed stator, laser welded housing, 304 stainless. It does not carry head, flow or horsepower. Those three still live only in the title. The site already filters on head and flow, so the data exists. It just never reaches the eye.
- V1, collection cards: head, flow and horsepower as three short lines under the name, plus stock status.
- V1+V2, product page: the same three numbers directly under the title, with the existing spec sheet linked beside them.
- Shorter product names on both, since the specs no longer have to live there.
"need a better technical comparison of the two pump options."Post-purchase survey, 29 June 2026, $1,694.65 order
Real product, real specs, taken from the live 220V collection. The product-page version repeats the same three lines under the title.
What this is built on
| Signal | Source |
|---|---|
| Choosing between models is a named friction. 17 of 161 post-purchase answers describe not being sure they picked the right product. "Haha just confused by all the different types and wanted to make sure this was the right one ..." | Post-purchase survey · 90 days |
| The site can already filter by head and flow, but not show them. The collection offers Head/Depth and Flow Rate as filters, then shows cards with neither. | Live site · captured 20 August |
| People reach the product cards and click the long titles. 94.46% reached the first row of products, and the three most-clicked elements were product titles. The comparison problem is present at the point of choice. | Microsoft Clarity · 1,065 desktop views, 30 days |
2
Why we're testing
A pump is not a finished purchase. It is the first item on a list the customer is still writing, and they know it. The drawer is where that doubt lands, because it is the first place they see the whole basket at once. Right now it shows them what they picked and says nothing about what is missing, so they either guess, or they stop and phone a rep, or they leave and come back later. The drawer even ends with the line "Real Person Customer Service (no AI)" and then gives no way to reach that person. The number sits back on the product page, one step earlier. The doubt arrives in the cart and the answer does not.
- V1, the parts row: a "Complete your set" row of three items, keyed to the pump already in the cart. One tap to add, without leaving the drawer.
- V1+V2, the expert line: "Unsure? Check with our experts" under the row, with the number the reps already answer.
- Fixed lists per pump, no app and no matching engine. Booster pumps get neither, because those buyers add nothing.
"trying to verify I had everything from rps needed before completing."Post-purchase survey, 17 August 2026, $2,539.98 order
"I spoke with a person on the phone to make sure this was the correct pump I needed before ordering. They were super helpful."Post-purchase survey, 14 July 2026, $1,263.80 order
Drawn over the live RPS cart drawer captured 21 August. The three parts shown are the actual top co-purchases for a Core Series well pump. The phone number is the one already used on product pages.
What this is built on
| Signal | Source |
|---|---|
| This exact surface already produced a win by making the gap visible. Items per order rose 18.2% while order value stayed flat at about $713. Buyers added more line items, not more expensive ones, the moment they could see the basket. | RPS04 · concluded July |
| Customers say the phone call is what gets the order right, unprompted, in reviews about the product. 23 of 736 published reviews describe calling before or during the purchase to check the choice. "Ordered a deep 1.5 hp pump, called in and got sizing help ..." Rep-created orders attach a second product 51% of the time against 31% on the website. | Judge.me · 736 reviews · Shopify · 59 days |
| The list is short, stable and specific to well pumps. Across 237 Core Series orders the same few parts recur, led by the One and Done Drop Pipe Kit on 20.7%. The 3HP booster, the highest-volume product on the site, attaches anything on 5% of its 135 orders and the 2HP booster on none of its 35, so boosters are left out. | Shopify · 1,196 orders, 59 days |
3
Why we're testing
Every product on the site shows the same rating and the same review count: 4.5 stars, 199 reviews. Open a pump page and the reviews underneath are about other products. On the 1.5 HP pump page today, two of the visible reviews are about a different pump end and a fountain motor, and two of them are people describing returns. The survey identifies one segment, about 8% of customers, whose main information source is reviews and who already feel negative. These are the pages they are reading.
- Show each product its own reviews and its own rating.
- Scoped to the bestsellers, not the whole catalogue. The top six revenue pages carry 61, 33, 16, 11, 10 and 9 reviews each.
- Anything under five reviews keeps the shop-wide rating, so no page looks emptier than it does today.
"I was going to buy today, but I wanted to take some time to read some reviews first. ..."Post-purchase survey, 9 June 2026, $854.99 order
Same widget, correct data. 86% of the reviews collected in the last 90 days already carry a product attribution.
What this is built on
| Signal | Source |
|---|---|
| There is enough review volume where it counts. 703 published reviews spread over 122 products. 36 products have five or more, and those pages carry half of all product revenue in the last 90 days. The long tail is thin, which is why the rollout stops at the bestsellers. | Judge.me · full account, pulled 21 August |
| The most hesitant segment reads reviews first. Skeptical Browsers are about 8% of customers, the only group with negative sentiment, and reviews dominate as their information source. | Segmentation survey · 101 customers |
| The May winner made the rating clickable and won. Conversion rose 36.8%. This restores that element's credibility rather than adding a new one. | RPS02 · concluded May |
4
Why we're testing
Search is one of the most-used routes on the homepage, and Shopify recorded 3,421 search queries in the last 30 days. Results were returned for the tracked terms, but only 1,490 searches led to a result click. The weak spots are commercially useful queries, not random dead ends: pressure tank, solar, well pumps and the exact SKU 13RPS05 all have high volume and low click-through. This is a relevance and result-presentation problem worth fixing before adding more homepage routes.
- Start with the four high-volume, low-click searches: pressure tank, solar, well pumps and 13RPS05.
- Use Shopify Search & Discovery to tune synonyms, product ordering and featured results for each query.
- Capture the current results for those four searches before briefing any result-page layout change.
"pressure tank" produced 36 searches and only 5 result clicks. "solar" produced 30 searches and 3 clicks.Shopify search data, last 30 days
This is a query-priority view, not a guessed redesign of the results page. The current results for these terms still need capturing before the build brief.
What this is built on
| Signal | Source |
|---|---|
| Search is a primary homepage route. On desktop the search input was the second most-clicked element after the pop-up close, with 275 clicks, plus 180 clicks on submit. On mobile the input ranked fourth with 367 taps. | Microsoft Clarity · 15,471 homepage views, 30 days |
| Only 43.6% of searches led to a result click. Shopify recorded 3,421 queries and 1,490 queries with clicks. The tracked terms returned results, so the issue is not simply empty pages. | Shopify search analytics · 30 days |
| High-intent terms are underperforming. Pressure tank clicked 5 of 36 times, solar 3 of 30, well pumps 7 of 30 and exact SKU 13RPS05 8 of 32. By contrast, one and done clicked 38 of 44 times. | Shopify search analytics · 30 days |
5
Why we're testing
The site answers price with a discount off list. But the alternative a buyer is actually weighing is not another website, it is phoning a local well company. The survey puts price fourth among decision drivers, behind service and convenience, and found free shipping outperforms both a $200 and a 25% discount. That points away from discounting and toward showing what the whole job costs the other way.
- A product-specific side-by-side on pump PDPs: that RPS pump vs. pump plus markup plus trip plus labour.
- Backed by real quotes, not estimates.
- Placed close to the product price and purchase area. Below the full specifications would put it beyond most visitors; the homepage would make it too generic.
The left column is real, from a live product page. The right column cannot be filled in until we have a genuine installer quote.
Recommended placement: pump product pages, directly after the purchase area, where the RPS price is specific and the customer is actively weighing the alternative. On the sampled PDP, only 40.23% of desktop visitors reached the lower trust section, so this comparison should not sit below the full specifications. The figures still need a genuine installer quote before build.
What this is built on
| Signal | Source |
|---|---|
| Price is the number one friction in the post-purchase survey, three cycles running. 34 of 161 answers, 21%. | Post-purchase survey · 90 days |
| But discounts are not what wins. The segmentation survey rated free shipping 4.3, ahead of both a $200 discount and 25% off at 4.1. It concludes price cuts may erode margin unnecessarily. | Segmentation survey · 101 customers |
| A savings callout already won here. Making the dollar saving explicit next to the price was part of the May winner, conversion up 36.8%. | RPS02 · concluded May |
6
Why we're testing
Customers say they are confused by the product types and struggle to find the right parts. They know what they are trying to do: run the house, water the crops, fill the stock tank, keep the pond alive. Every route into the catalogue asks them for something else first, horsepower, voltage, wire type, motor diameter, GPM. Crutchfield solves the same problem for car audio by leading its filters with what the thing is for, then the specs underneath.
- Four entry tiles: household, produce, livestock, pond and fountain.
- The same four as the first filter group on collection pages.
- Spec filters stay exactly where they are, just below.
Nothing is removed. A layer the customer already thinks in is added above the layer they have to learn.
What this is built on
| Signal | Source |
|---|---|
| Four jobs cover almost everyone. Household 62%, produce irrigation 48%, livestock 25%, ponds and fountains 21%. Respondents could pick more than one, so these overlap and total more than 100%. The survey recommends restructuring the homepage around these four. | Segmentation survey · 101 customers |
| Collection visitors actively use the technical filters. Filter checkboxes account for several of the ten most-clicked elements on the sampled 220V collection. The opportunity is to add a plain-language first step, not remove the filters. | Microsoft Clarity · 1,065 desktop views, 30 days |
| Crutchfield leads its filters this way. On receivers the first filter group is "Recommended For", with the technical facets underneath, each carrying an info icon. | Crutchfield · captured 21 August |
Not tests
Four things the data surfaced that no A/B test will solve. They are here because they cost money or trust today, and most take minutes rather than a slot.
Payment methods are failing at the moment people pay
18 of 161 post-purchase answers name a problem at the payment step, and the pattern has now held for two cycles. PayPal links that never connect, Apple Pay locking out a card switch, an authenticator loop, a missing pay-over-time option. Separately, 67 of 77 abandoned checkouts in the last 30 days reached the payment step, on a median basket of $961.83. The survey found abandonment is driven by confidence rather than price, so a working PayPal button matters more than a discount email.
"Unable to get Paypal link to work. Neither link from your website worked properly and the PP logo came up but never connected. I ended up providing the CC number directly on your website."Post-purchase survey, 10 July 2026
The sizing calculator popup is driving complaints and unwanted calls
One customer describes it precisely: it repeats, completing it does not dismiss it, and finishing it triggered sales texts and phone calls they did not ask for. There are currently six separate entry points to the sizing wizard across the header, the nav, a homepage popup, a homepage banner, a homepage tile and a collection banner. Worth checking today.
"Website keeps popping a sizing calculator that is annoying. Clicked through it to see if that would make it go away. It didn't, but now I am getting text and phone call attempts from RPS."Post-purchase survey, 29 July 2026, $358.86 order
The accreditations are not on the site
BBB A+ accreditation, the 4.9 Google rating across 106 reviews, and the Water Assurance Guarantee badge do not appear anywhere a customer sees them. Crutchfield states four specific trust numbers on both its homepage and every product page. The most hesitant segment, about 8% of customers, is the one that needs exactly this before committing.
Show the agricultural tax exemption before payment
RPS already has an agricultural tax exemption process, but the link sits in the footer. Two customers raised the tax unprompted in the last month, including one who had an exemption number ready. This is an existing process that needs to be visible in the cart or checkout, not an A/B test.
"The sales tax. This is for tax-free agricultural use. I have my number and card available."Post-purchase survey, 5 August 2026, $811.29 order