Why that one pool keeps going cloudy

Okay so I want to start with a pool that doesn't make sense. And I'm guessing some of you have one of these right now, or you had one last season and you never actually figured out what was going on with it.

Here's the pool. It goes cloudy. Not dramatically, not a green swamp, just, dull. Every few weeks. And you pull the log and the log says the chemistry's fine. Chlorine's there, pH is fine, alkalinity's fine. So you shock it, it clears up, everybody's happy, and three weeks later it's dull again. After the third or fourth round of that you start to get this feeling, like, either I'm going crazy or this pool is haunted.

It's not haunted. What's happening, almost always, is that, um, that the numbers in that log were never real numbers. They were, um, numbers-shaped things. Something got written in the box, but the thing in the box wasn't a measurement of that water. And that's the whole episode, honestly. Everything we published this week -- salt cells, filters, shock, algaecide, how to go get commercial accounts -- all of it sits on one idea, which is that a number is only worth whatever the method behind it was worth.

So let's start at the absolute bottom. Which is where you put the bottle in the water.

We ran a piece this week on how to actually test pool water on a route, and when I first saw it on the schedule I thought, okay, this is the most basic thing we could possibly write about. Then I read it and thought, no, this is the thing everybody's been doing slightly wrong for eleven years and nobody's ever told them.

So. First rule, and it's the big one: you test before you touch anything. Like, first. The sample is your first act at that pool. Before you brush, before you empty a basket, before you put anything in the water. Because whatever you do to the pool first is now in your sample. Dose acid and then test, you're not testing the pool, you're testing the acid. Brush the walls and then test, you've just put whatever was on the walls into suspension. And people get this backwards for workflow reasons -- you get there, you start working, and testing feels like a thing you do while other stuff is happening. But it's the measurement, and a measurement has to come before you change the thing you're measuring. Sounds obvious when I say it out loud. It is not obvious at 2pm on your twenty-eighth stop.

Second one, and this is the one that I think surprises people most: go deep. Elbow-deep. Push the bottle in upside down to about eighteen inches, then flip it upright down there and let it fill. Not a scoop off the top. Because the top few inches of a pool are not the pool -- that layer's been sitting in the sun, it's warmer, it's chemically weathered, it's got whatever blew in on it. On a hot afternoon it can read meaningfully different from the water underneath. So if you're skimming the surface with your bottle, which everybody's done because it's easier and you don't have to kneel down, you're reading a layer nobody's swimming in.

And then where you stand. At least three feet from any return fitting, and don't sample at the skimmer throat. Water coming out of a return has just been through the equipment, and on a pool with an in-floor system it's just been mixed with whatever the equipment last pushed through. So you're reading the equipment, not the pool. Deep end, or the middle of a long wall.

Then the order you run the tests in. Free chlorine first, always, because chlorine's the most time-sensitive reading on the panel -- it degrades in a warm sample bottle in direct sun. Run it last, after the bottle's been sitting on the coping while you did pH and alkalinity, and you'll read low. And here's what gets me about that: it doesn't read low in a way that looks wrong. It reads low in a way that looks like a pool that needs chlorine. So you dose it, and now you've overdosed a healthy pool because your sample cooked. So chlorine, then pH, then total alkalinity, then everything else that doesn't move on a ninety-second timescale.

Okay, two more and I'll move on. Check the date on what you're testing with. Reagents drift as they age, strips drift as they age, and neither one fails in a way you can see. There's no warning light. The bottle just quietly starts producing numbers that aren't true. And a fresh bottle costs less than one wasted return trip to a pool that was fine the whole time.

And the last one, and this is, I mean, this is the one. Write the numbers down at the pool. While the bottle's still in your hand. Not at the next stop, not in the truck at the end of the day. And this is the step that gets skipped more than any other step in this entire job, and it does the most damage, because -- okay, you're reconstructing four readings across forty pools from memory at 6pm. You're not doing that. Nobody's doing that. What actually happens is you write down what the pool usually reads. Which is how a log fills up with numbers that describe your expectations instead of the water. Which brings us all the way back to my haunted pool.

Right, so, this is where the second piece of the week comes in and honestly this is the one I'd send to anybody running more than one truck. It was on test kits versus test strips, and the framing I really liked is that this argument never resolves when you ask it as a yes or no. Are strips accurate enough? Well, for what.

Split it by reading, though, and it gets easy. For pH, strips are genuinely fine -- a fresh strip lands within about two tenths of a unit, and you're deciding whether to add acid, not chasing a hundredth. For "does this pool have any chlorine in it at all," strips are great. Ten seconds and you know. On a forty-stop day that's a real tool.

The trouble starts the second the reading sets a dose. Strips read free chlorine in printed bands -- zero, half, one, three, five, ten -- so every actual value gets rounded to the nearest block of color on the label. A pool sitting at 4 ppm reads as 3, or as 5, depending on the light and the strip and who's holding it up. That's a sixty percent spread on the number you're about to dose against.

And that gets genuinely bad once cyanuric acid's in the picture, because the chlorine you need isn't a fixed number. You hold free chlorine at roughly seven and a half percent of the CYA reading, so a pool at CYA 60 needs about 4.5 ppm. The strip's nearest bands are 3 and 5. One of those is under-sanitized and heading for algae, the other's fine, and the strip physically cannot tell you which one you've got.

But here's the part that I think actually costs money, and it's, you know, it's not about accuracy at all. It's repeatability across a crew.

So the piece had this example that I keep, like, coming back to. Operator with three techs, 210 pools, Henderson and southeast Las Vegas. Two of the techs carry strips, one carries a drop kit. There's a pool that reads CYA 50 on a Tuesday, and then 90 the following Tuesday. Nobody added stabilizer. No water was replaced. And nothing happened to that pool -- the first number came off a strip pad and the second one came off a titration, and that forty-point jump is just, that's the gap between two methods. That's it.

And the cost isn't the wrong number. It's the month of decisions made on it. At CYA 50 the sanitizing target's around 3.75 ppm. At CYA 90 it's closer to 6.75. So one tech's dosing to a target nearly three full points below where the other tech thinks it should be, on the same pool, and both of them are following the rule correctly. The pool drifts, algae shows up in week three, and the log doesn't explain anything, because the log looks like the readings were taken properly.

That's a step in the history that never happened. It's an artifact of equipment, not a change in the water. And I think that's the thing to really sit with -- a per-pool reading history is only interpretable if the readings were produced the same way. Otherwise you're reading a chart where somebody swapped the ruler halfway through.

So the answer isn't strips bad, kits good. It's one method per reading, across the whole crew, written down. Most service businesses land on a mixed setup and it's cheaper than people assume -- one drop kit or a photometer per truck, strips in the door pocket for the quick pass, and a written rule about which reading uses which. And if you only upgrade one, upgrade cyanuric acid, because the CYA pad on a strip lands within about thirty ppm of the truth, and CYA sets the target for the reading you check every single visit. The real test, you add reagent till a black dot at the bottom of the tube disappears. Ten ppm steps, doesn't care about the light or your eyesight. Takes about thirty seconds longer than a strip, and it's the best thirty seconds in the panel.

Okay. So that's producing a good number. Now I want to talk about the numbers that are bad in a way you can't see on any single visit, because this is where two more of this week's pieces live and I think this is the most underrated thing we published all week.

We ran a comparison of the three shocks -- cal-hypo, dichlor, liquid. And the framing is: all three deliver hypochlorous acid, all three will clear a pool. What separates them is the passenger each one brings along.

Cal-hypo is 65 to 73 percent available chlorine, it's the cheapest chlorine per pound, and it survives a hot truck bed better than anything else. Every dose also puts calcium in the water. Dichlor is stabilized granular, 56 to 62 percent, dissolves fast, close to pH neutral, super convenient to carry and measure. Stabilized means it carries cyanuric acid, so every dose adds some. And liquid chlorine, sodium hypochlorite, 10 to 12 and a half percent at professional strength -- adds no calcium and no CYA, which on a route is the entire argument for it. And it's heavy, it eats bed space, and it starts losing strength the day it's made.

Now, one dose of any of those is harmless. The problem is, um, the problem is the fiftieth.

Dichlor adds roughly 0.9 ppm of cyanuric acid for every 1 ppm of free chlorine it delivers. So shock a pool to 10 ppm six times over a season and you've put something like 54 ppm of CYA into water that probably started the year around 40. That pool's now stabilizer-heavy, the chlorine you're paying for is doing less work, and -- this is the part -- nothing on your truck removes cyanuric acid. Almost every other reading you correct with a dose. CYA you correct by replacing water, on a pool where the customer pays for the water and in a drought market might not be allowed to use much of it.

And cal-hypo does the same thing to calcium hardness. At 73 percent it adds around 0.7 to 0.8 ppm of hardness per 1 ppm of free chlorine, so those same six shocks put about 45 ppm of calcium in over a season. On soft fill water, genuinely fine. On Sunbelt fill water already arriving at 300 to 400 ppm, that's a countdown.

And the example in the piece is an operator running 160 pools out around Mesa and Apache Junction, fill water at 380 ppm hardness. He standardized on cal-hypo years ago for a completely honest reason, which is that it survives a July truck bed. Two seasons later, roughly a third of his pools are reading over 600 ppm calcium and scaling on the tile, and the fix on each one is a partial drain the customer never budgeted for.

And what I love about that -- love is the wrong word -- what, I mean, what gets me about that is nothing went wrong on any single visit. The product was correct fifty times in a row and the total was still wrong. There's no bad day in that story. There's no tech who messed up. It's just arithmetic that only exists in the trend.

Which is the whole reason the reading history matters more than the last reading. Look at one visit and that pool's fine every time. Look at a season of its readings on the pool's record and you watch the hardness number walk upward, while it's still a dosing decision instead of a drain. Drift is invisible at the visit level and obvious at the history level, and you only get the history if somebody wrote the number down while the bottle was still in their hand.

So the practical read on shock is, I think it's: liquid's the sane default for weekly dosing, because it's the one that adds neither. Hold cal-hypo back for algae recovery and for pools that are genuinely low on calcium. And the reason routes drift away from liquid isn't chemistry, it's logistics. Sodium hypochlorite decomposes on its own, and heat roughly doubles the rate for every fifteen to eighteen degrees. At 77 degrees a 12.5 percent drum loses around 10 percent of its available chlorine in a month. Parked in a truck bed at 95-plus through a Phoenix August, that same month can cost you 25 to 30 percent. And you won't notice that as a bad batch. You'll notice it as pools that suddenly need more product than they used to, which reads like a chemistry problem and is actually a storage problem. So the fix is a shorter restock cycle, not a different product -- buy two to four weeks at a time, keep it out of the sun and off a hot metal bed, rotate stock instead of stacking new behind old.

Okay, and while we're on the subject of things that accumulate where you can't see them, let me wander sideways for a minute, because we also did a piece on algaecide and it belongs right here.

So the first thing, and this one's a genuine correction for a lot of people: algaecide does not kill a bloom. Chlorine kills algae. Algaecide prevents algae from establishing, or it backs up a treatment. And treating it like a killer costs you a visit, because you dump a bottle in a green pool, you come back a week later, and it's still green, and now you've lost seven days and a bottle.

So you pick algaecide by side effect, not by strength. Three classes. Polyquat is the polymer one -- it keeps algae from establishing, clarifies the water a little, no stain, no foam, doesn't add chlorine demand. Highest cost, roughly 25 to 40 bucks a quart, maintenance dose usually 2 to 4 ounces per 10,000 gallons. That's the safe default.

Copper works. Genuinely works, and it persists long after the chlorine's burned off. That persistence is the entire problem, because copper doesn't leave on its own, and above roughly 0.2 to 0.3 ppm it starts showing up as stain. On plaster. In blonde hair. Which is a phone call you really don't want.

Quats are the cheapest bottle and the weakest preventative. It's a surfactant, so it foams when you overdose it, and it consumes chlorine rather than sparing it. Oh, and sodium bromide gets sold right alongside these and isn't really an algaecide in the same sense -- it converts to bromine when you shock, and it exists almost entirely for black algae.

But look at what copper has in common with cal-hypo and dichlor. It's a thing where the individual decision is defensible and the accumulation is the problem, and the accumulation is only visible in the pool's own history. Same shape. Different chemical. And a well-chlorinated pool mostly needs none of it, which is worth saying out loud before anybody standardizes a route on a preventative dose.

Alright. Let me get out of the water and go stand at the equipment pad, because there's a version of this exact same idea living over there too.

Filter cleaning. And the piece on this made a point I hadn't heard put quite this way: filter cleaning is the maintenance job most likely to be done on the wrong schedule, and the reason is that homeowner advice puts it on a calendar. Every month, every quarter, at opening and closing. Because a calendar is the only trigger somebody with one pool can actually act on.

On a route you've got a better trigger and it's sitting right there on the pad. You clean when pressure is running 8 to 10 psi over what that specific pool reads clean. Not over some generic number. Over that pool's own clean baseline. Which means the whole thing only works if somebody wrote the baseline down.

And here's the bit that I think is the actual tip in the entire piece: read and write down the pressure after the clean, not just before. Because the after number is next visit's baseline. And it's the only reading that makes the next pressure delta mean anything. Everybody records the high pressure, because the high pressure is why you're doing the job. Almost nobody records the number after, and that's the one that has the value.

The methods themselves are quick. Cartridge: pump off, release the air relief, open the housing, lift the element out, hose it top to bottom at an angle that opens each pleat, two or three times around. And never backwash a cartridge -- there's no backwash position on that plumbing, and running it in reverse drives the debris deeper into the pleat.

DE gets backwashed rather than opened. Two to three minutes till the sight glass runs clear, rinse thirty seconds to reseat the grids, then recharge with fresh DE through the skimmer at the manufacturer's dose for that grid area. The recharge is the part people skip, and backwashing a DE filter without recharging it is the single most common DE mistake -- it leaves the grids bare and the water hazy.

Sand's the fastest. Backwash till the glass runs clear, usually two to three minutes, rinse twenty or thirty seconds, back to filter. Fastest of the three and the least thorough, which is why sand wants a chemical deep clean once a year rather than just a seasonal rinse.

And a loaded filter is the most common reason water goes hazy on a pool whose chemistry reads fine. Which, again, my haunted pool from the top of the show -- sometimes it's not the chemistry at all, and you'd know that if you had a clean-pressure number to compare against.

Salt cells are, honestly, the same story wearing a different hat. So the rule there is: look into the cell every visit, and clean it only when you can actually see white scale on the plates. Because acid cleaning erodes the plate coating and shortens the cell's life. So cleaning a cell that didn't need it is actively destroying it faster. That's a weird one to get your head around -- the maintenance is the damage.

Procedure's short. So, kill power at the control board before you touch anything, then unthread the cell -- doing that with the system live is how you get a shock and a cracked union in the same afternoon. Look at the plates in daylight before you commit, because light scale sometimes just rinses off with a hose, and if the hose clears it you're done, you just saved that cell an acid bath it didn't need. If it does need it: one part muriatic acid to four parts water, acid into water, never the other way around, eye protection, open air. Pouring water into acid is the one that splashes. Soak till the bubbling stops, usually ten to twenty minutes, and when it goes quiet the calcium's off and every minute after that is acid working on the plate coating instead. Rinse, reinstall, restore power, and confirm the cell's reporting output before you leave the pad.

And then log that you cleaned it. On that pool's record. Because the date of the last cleaning is what turns next visit's decision into a two-second call instead of a guess.

And -- okay, here's where it loops back around. A cell that scales over and over again does not have a cleaning problem. It has a calcium problem. Which is the exact thing we were talking about twenty minutes ago with cal-hypo on hard fill water. Same pool, same number, showing up as a different symptom on a different piece of equipment. And you'd catch that by looking at the hardness trend on that pool's history, not by cleaning the cell a fourth time.

Last one on the equipment side, and this one's kind of a fun diagnostic. Sand in the pool. There's grit on the floor, the customer's annoyed, and the question is which part failed.

First, rule out that it's not filter sand at all. Plaster dust reads as grit and it isn't coming from your filter. Do that before you buy any parts, because otherwise you're replacing a lateral to fix a plaster problem.

Then, three real causes, and this is the good part -- when the sand shows up tells you which one it is. A cracked or unseated lateral -- those are brittle plastic fingers living under a quarter ton of wet sand, and they crack from age, from freeze damage, or from somebody standing on them during a media change. One split lateral feeds a steady trickle into the return line. A broken standpipe, that's the vertical tube from the multiport down to the lateral hub, and if it cracks, sand pours straight into the pipe carrying filtered water back to the pool. Or a failed spider gasket in the multiport, which isn't a filter fault at all -- when that gasket hardens or tears, flow crosses between ports, so water carrying fluidized sand reaches the return during a backwash instead of going to waste.

And what it's almost never: too much sand in the tank, or sand that's simply old. Overfilling causes carryover into the multiport, not a fan of grit on the floor.

So that diagnosis is pattern recognition over time. Same muscle. You're reading a history, not a snapshot.

Okay so. Two pieces left and this is the turn, because everything I've said so far has been about craft, and these last two are about what the craft is worth.

Adding repair services. And every pool cleaning business gets this call, right, like, every single one. A pump's screaming, a heater won't fire, and the customer asks the person they already trust with the pool. And most operators refer it out and then watch a contractor collect six hundred dollars for two hours of work and think, why am I not doing that.

And the numbers are kind of brutal to look at. A $165 monthly account bills just under two grand a year, and chemicals and fuel and roughly fifty hours of labor come out of that before anything's left. One pump replacement on that same pool bills seven hundred to sixteen hundred in an afternoon, parts marked up 25 to 50 percent over landed cost, labor at 85 to 150 an hour. So a single repair invoice can carry as much gross margin as a full year of cleaning that pool.

But it's also the fastest way to lose control of a service day, because a ninety-minute repair dropped inside a twenty-five-minute stop wrecks everything behind it. So you stage it. Three separate decisions, not one. Stage one is diagnose and quote while you refer the actual work out -- a diagnostic fee, 75 to 150 bucks for a pool you don't service, covers the trip and the thirty minutes of finding the fault, and it filters the calls that were never going to become work. Almost nobody charges for it and it's basically free money. Stage two is simple part swaps, the failures you already see weekly -- a motor that won't start, a cell reading low, a collapsed cartridge. Roughly 70 percent of the repair calls on a residential route land there. And stage three, installing equipment, that's a different business with a different insurance policy, and in a lot of states a contractor license question that isn't a business decision at all. Don't walk into stage three because one customer asked about a heater.

And here's why this belongs in this episode. The reason your customers call you and not a repair company is that you're the person who's been at that pool every week and has the equipment history. That's the whole advantage. And it's exactly what we've been building this entire episode -- the pressure baselines, the cell cleaning dates, the hardness trend. That's the equipment history. And it is not in your head if you've got sixty pools. It's on the record or it doesn't exist.

And the margin on repair work leaks in one specific place, which is between the yes and the invoice. A number said out loud at the pad becomes the price, the parts receipt goes missing, and then the job gets billed three weeks later at whatever anybody remembers. So the estimate has to be its own record, written before you touch the equipment, and then it turns into the invoice. Same discipline. Different form.

And then the last piece of the week, which is going after your first commercial accounts, and this is the one that made me want to build the episode this way.

Because commercial is a different product, not a bigger pool. There's a compliance floor you clear before anybody talks price -- a certified operator on the account, general liability raised to what commercial buyers actually ask for, the entity added as an additional insured. That step takes the longest and nobody can shortcut it for you.

But then there's this line in the piece, and when I read it I thought, okay, this is the episode. A property manager is going to ask to see what a month of your documentation looks like before they hand you a pool. And -- records you can export and email in a minute are a selling point. A bucket of paper test slips is a disqualifier.

That's it. That's, that's the payoff for everything else. Where you dipped the bottle, whether the crew was on one method, whether somebody wrote the after-pressure down, whether the readings went in at the pool or got reconstructed in the truck -- all of that turns into a document that either wins you a commercial contract or loses you one. The discipline isn't for its own sake and it isn't so your records look nice. It's inventory. It's the thing you hand across the table.

And the rest of the commercial playbook is short. Win one reference account first, a small one, an eight-unit apartment building, a little gym -- it's worth more as a reference than as revenue, because most managers won't be your first, and one live account ends that objection permanently. Approach the manager, not the owner, because these pools get contracted by property management companies and HOA boards, so find who actually holds the contract before you write anything. Time it to the renewal, because they mostly run annually and get re-bid on a schedule, and showing up in month three of a twelve-month term gets you a polite file-away. And price for the real scope -- daily testing, weekend coverage, logbook upkeep, net-30 terms, those are all costs, and a commercial pool priced off your residential rate loses money quietly for a year. The net-30 part is the one that sinks people, because you're floating that payroll.

So. Back to the haunted pool.

The reason it kept going cloudy is a measurement problem, not a chemistry problem. Either the sample came off the surface. Or it came from right next to a return. Or the chlorine got tested last and read low because the bottle cooked. Or the CYA came off a strip and the real target was three points higher than anybody thought. Or -- honestly the most common one -- the filter was loaded, the chemistry was never the issue, and nobody had a clean-pressure baseline to check against.

None of those is a hard problem. Every one of them is a habit. Where you put the bottle, what order you run the tests in, whether the crew's on the same method, and whether the number goes down at the pool or gets remembered later.

And what I want to leave people with is that none of this is about being fussy. Fifty visits from now, that pool's hardness number is either a dosing decision or a partial drain, and which one it is got decided by whether the trend was legible. And two years from now, when you're across the table from a property manager who wants to see a month of documentation, what you're handing them is just, it's the accumulated output of somebody kneeling down and pushing a bottle eighteen inches under the water before they touched anything.

That's the whole thing. Alright. That's the week. I'll see you next one.