THE PILLAR
Considered purchases: why buyers hesitate, and what to do about it.
A considered purchase is one people research before they commit. The decision takes days, not minutes; it involves comparison, configuration and often another person; and hesitation is the normal state of the journey rather than a defect in it.
A considered purchase sits roughly in the €300–€3,000 band, involves comparison and configuration before checkout, and is bought once every few years rather than weekly. The buyer is neither browsing nor certain. They are doing careful work, and they can be lost at any step of it without ever telling you why.
This page is the argument Yesward is built on: that these journeys are the ones where restraint has commercial value, because the buyer's attention is already committed to the decision and interrupting it costs more than it does on an impulse purchase.
WHY THE DISTINCTION MATTERS
Tactics that work on a €40 basket are actively harmful on a €1,800 decision.
| Impulse / low consideration | Considered purchase | |
|---|---|---|
| Decision window | Minutes, one session | Days to weeks, several sessions and devices |
| What abandonment means | Mild interest that faded | A specific unanswered question, usually a knowable one |
| Effect of urgency tactics | Often works — the stakes are low enough to gamble | Reads as pressure on a decision the buyer is being careful about |
| Effect of a discount | Tips a marginal basket over the line | Often pays people who would have bought anyway, and rarely answers the real question |
| Value of one recovery | Cents of margin | Enough that one recovered contract can pay for months of platform |
| Right default | Ask for the sale | Get out of the way, and answer only what is actually blocking them |
This is why we do not think of Yesward as an ecommerce conversion tool with a considered-purchase mode. The economics of restraint only work when a single recovered decision is worth real money — which is exactly this band.
THE JOURNEY
Four steps. Hesitation clusters at the seams.
comparison → configuration → price → checkout / application
STEP 01 · COMPARISON
Which of these, and why
Three tabs open, a spec table half understood, and no basis on which to separate two options that both look fine.
Dominant friction: choice
SIGNALS: A↔B LOOPS · COMPARE PAGE RETURNS
STEP 02 · CONFIGURATION
Will it actually work for me
Dimensions, dates, address, eligibility, compatibility. The step where a buyer discovers a constraint nobody warned them about.
Dominant friction: coverage
SIGNALS: REPEAT CHECKS · CHANGED INPUTS
STEP 03 · PRICE
What will this really cost
Not “is it too expensive” but “what is the total, over what period, with what added”. Usually a clarity problem wearing a price costume.
Dominant friction: price, then commitment
SIGNALS: FEE FAQ · CALCULATOR RE-RUNS
STEP 04 · CHECKOUT / APPLICATION
Am I sure about committing
Contract length, cancellation, what happens if it goes wrong, whether this company can be trusted with a direct debit.
Dominant friction: commitment, trust
SIGNALS: TERMS PAGE · TERM TOGGLES · STALL
Analytics can tell you which step people leave from. It cannot tell you which of the four questions above they were stuck on — and the answer determines whether the right response is a sentence, a question, or nothing at all.
THE VOCABULARY
Five frictions. Each one has a different cheapest answer.
Choice friction
Two or more options look equally valid and nothing on the page separates them.
Cheapest answer: the one distinction that actually decides it — usually not price.
TYPICAL LEVEL · L1–L2
Price friction
The total cost is unclear, or the shape of it over time is unclear.
Cheapest answer: information. A discount is a last resort you switch on deliberately.
TYPICAL LEVEL · L1–L3
Commitment friction
The term, the lock-in, the exit. Very common in subscription and service purchases.
Cheapest answer: one question about the buyer's horizon, then the matching term.
TYPICAL LEVEL · L3 · THE WORKED EXAMPLE
Trust friction
Doubt about the company, the promise, or what happens when something goes wrong.
Cheapest answer: the specific promise in writing. Invented social proof makes it worse.
TYPICAL LEVEL · L1
Coverage friction
Eligibility, availability, dates, address, compatibility — a hard constraint, not a doubt.
Cheapest answer: often none. If the answer is no, the honest move is to say so early.
TYPICAL LEVEL · L0 OR L5
Note what is missing from this list: “not interested”, “just browsing”, “price too high”. Those are conclusions, not frictions — nothing on your website can resolve them, so the correct response is silence, and a receipt recording that we decided so.
WHERE THIS SHOWS UP
Four verticals where hesitation is the whole game.
Chosen for structural fit rather than sector glamour: comparison behaviour, configuration steps, real acquisition spend, and margin data that exists server-side.
01 · HOME & FURNITURE
Will it fit, and will it last
High order value, visual products, specification and material questions, delivery windows, financing. Comparison behaviour is intense and mostly invisible in analytics.
DOMINANT: CHOICE · COVERAGE
Case-study shape: reduce product comparison abandonment without discounting.
02 · TRAVEL
Dates, rooms, refunds
€500–€3,000 bookings, multiple travellers, cancellation conditions, room and date selection, and enormous acquisition spend behind every session.
DOMINANT: COMMITMENT · CHOICE
Case-study shape: reduce booking hesitation at the date and refund step.
03 · INSURANCE & CONSUMER FINANCE
Am I eligible, and what is excluded
Friction around cost, eligibility, commitment and complexity — the strongest product fit of the four. The trade-off is a slower sales cycle into regulated organisations.
DOMINANT: TRUST · COVERAGE · PRICE
Case-study shape: reduce quote abandonment mid-application.
04 · ENERGY, TELECOM & MOBILITY
Which tariff, for how long
Tariffs, contract terms, hardware, installation, financing and comparison sites. Our first beachhead, and the journey the demo account is modelled on.
DOMINANT: COMMITMENT · PRICE
Case-study shape: reduce tariff comparison abandonment. See the test site ↗
THE RESPONSE
Match the smallest response to the friction you can actually evidence.
Every row below is a decision the engine makes thousands of times a day. The right-hand column is the part most tools skip: what we deliberately do not do, and why.
The Intervention Ladder in full →| Evidence says | Smallest useful response | What we do not do |
|---|---|---|
| Weak or ambiguous signal | Nothing. L0, with a receipt. | Guess, and call the guess personalisation |
| Choice friction | The one distinction that separates the options, in page flow | Recommend the highest-margin option and call it help |
| Price friction | Total cost made explicit, over the period the buyer is thinking in | Open with a discount. Off by default, policy-gated |
| Commitment friction | One question about their horizon, then the term that matches | Open a chat and hope a conversation happens |
| Trust friction | The specific promise, in writing, where the doubt occurred | Badges, counters, and “2 people are viewing this” |
| Coverage friction, answer is no | Say so early and stop spending their time | Capture the lead anyway |
| Buyer already progressing | Nothing. Especially nothing. | Congratulate ourselves on 96% personalisation coverage |
MEASUREMENT
Long journeys make attribution easy to fool.
A considered purchase spans sessions and devices, which means a large share of buyers convert some time after any intervention you made. Last-touch logic will happily credit your nudge with sales it had nothing to do with.
The only defence is a control group that persists as long as the programme does — and a reported unit that cannot be inflated by giving margin away.
The measurement maths → · Three ways NBA results get overstated →
FOUR METRICS WORTH THE ARGUMENT
Incremental gross profit
Treated minus permanent holdout, net of discount, model spend and fees. The headline number, with its interval.
Qualified outcome rate
Share of eligible sessions reaching a genuinely valuable outcome — not clicks, not conversations started.
Cost of speaking
Dismissals, rage clicks, bounce delta and suppressed sessions, reported next to the lift rather than in an appendix.
Restraint rate
Share of sessions deliberately left alone. A number most vendors would rather you did not ask for.
RESEARCH · IN PROGRESS
The Considered Purchase Friction Report
Every deployment observes friction. Aggregated anonymously across accounts, that becomes something no competitor can write from a keyword tool: which behavioural patterns indicate which friction, which responses resolve it, how often silence wins, and what each treatment was worth.
Not published yet, and we will not pretend otherwise. The first edition needs enough sessions across enough accounts to say anything defensible.
PLANNED CUTS
Founding pilots contribute anonymised aggregates and get the first edition before it is public.
How a founding pilot works →FREQUENTLY ASKED
Considered purchases, briefly.
What is a considered purchase?
A purchase people research before committing — typically €300–€3,000, involving comparison and configuration, bought rarely enough that the buyer has no habit to fall back on.
Why do buyers hesitate on them?
Because a specific question is unanswered: which option, will it work here, what is the total cost, what am I committing to, can I trust this company. Hesitation is rational, not a lack of interest.
Is a discount the fastest fix?
It is the fastest way to raise conversion rate and often the fastest way to lower profit, because most of it goes to people who would have bought anyway. It rarely answers the actual question.
How is this different from cart abandonment?
Cart abandonment is one symptom at one step. On considered purchases most loss happens earlier — during comparison and configuration — where there is no cart to recover and no email to send.
Does this work for B2B?
Where the purchase completes online at this value, yes. Enterprise deals with a sales cycle and a committee are a different problem, and we would rather say so than stretch the claim.
Where should I start?
One journey, one outcome, two weeks in shadow mode. You get the friction distribution for your own traffic before anything is shown to a visitor. The pilot →
PAGES IN THIS CLUSTER
/solutions/
Six problem pages
Checkout, quote, booking, comparison, friction, anonymous visitors
/company/methodology/
The Intervention Ladder
Seven levels, five frictions, and the words we use for them
/decisioning/no-action/
When no action wins
The essay, plus nine questions it answers
/compare/
Compare
AdaptiveCX, Wisepops, Pathmonk, Optimizely
Get the friction distribution for your own journey.
Two weeks in shadow mode tells you which of the five frictions your buyers actually hit, and how often the answer is nothing.