CRO · Research · Experiments
Conversion rate optimisation agency that builds the winning changes
In shortWeDevelop is a Romanian software company that runs conversion rate optimisation for online shops and lead-generating websites: funnel research, user testing, A/B experiments and checkout work. Because engineers sit with the marketers, proven changes are built into the site instead of staying in a testing tool. It fits businesses with steady traffic that leaks somewhere identifiable, while low-traffic sites get research-led fixes rather than underpowered tests.
At a glance
- Provider
- WeDevelop, a software company from Romania
- Clients
- Romania, the EU, the UK and North America
- Languages
- English and Romanian
- Invoicing
- EUR, GBP or RON
- Pricing
- Quoted after scoping, from your brief
- Contact
- office@wedevelop.ro
01Is this for you
When a CRO agency is the right spend
CRO works on visitors you already have, so it needs some volume and a site that can be changed. These are the usual signs.
- Paid or organic traffic keeps arriving, yet orders or enquiries stay flat.
- Analytics shows a clear drop at one step, such as delivery, payment or a long form.
- Mobile visitors behave very differently from desktop visitors.
- Someone has proposed a redesign, and nobody can say what it would fix.
- Traffic keeps getting more expensive and you want existing visitors to work harder.
- Not a fit yet: analytics cannot be trusted, or the real obstacle is your price or your offer rather than the page. Testing button placement will not fix either.
02What you get
What a CRO engagement can include
You can stop after the audit. The later parts only make sense if the research finds something worth acting on.
Conversion audit
Funnel data, recordings, forms and a heuristic review, ending in a backlog ranked by expected impact and effort.
Checkout and form work
Field-level drop-off, payment and delivery options, error messages and mobile behaviour, where shops tend to lose the most money.
Experiment programme
Tests designed, built and analysed, with sample sizes set beforehand and losing variants reported alongside the winners.
Implementation
Proven changes built properly into your site by the same team, or specified for your developers.
03Which option
A/B testing, research-led fixes or a redesign
The right route depends mostly on how many conversions you get and how much risk the business can carry. As a rough guide, A/B testing becomes meaningful at a few hundred conversions per month per variant.
| Route | Best when | Trade-off |
|---|---|---|
| A/B testing programme | Key pages reach that conversion volume and you want proof of what worked. | Each test takes time, and many will lose. That is still useful, although nobody enjoys paying for it. |
| Research-led fixes | Volume is too low to test properly, but user sessions and funnel data show obvious problems. | You compare before and after rather than running a controlled experiment, so other changes can blur the result. |
| Targeted rebuild | One template, such as checkout or the product page, is the proven weak point. | Development cost concentrated in one area, which needs a careful release and close measurement. |
| Full redesign | The platform or brand has to change anyway, or research shows the design itself is the problem. | Everything changes at once, so you learn little about what helped, and the risk is higher. |
We do not default to a redesign. If one is unavoidable, we record a baseline on the old site first and release in stages where the platform allows it.
04Cost
What decides the cost of a CRO programme
Two sites of the same size can need very different programmes, depending on what research is possible and how easily the site can be changed.
| Factor | Why it matters |
|---|---|
| Traffic and conversion volume | Decides whether experiments are possible at all, and how long each one has to run. |
| Research depth | Analytics and recordings are one level, while moderated sessions with recruited users are more work. |
| Who can change the site | Direct deployment keeps the cycle short. Locked or heavily customised platforms limit what can be tested. |
| Checkout complexity | Payment methods, delivery options and B2B rules multiply the cases to build and check. |
| Measurement condition | Funnel events have to be right first, so faulty instrumentation means repair work before research starts. |
| Testing and recording tools | Licences, where a tool requires one, sit outside our work and depend on your choice and your traffic. |
The audit is priced as a standalone job. An experimentation programme runs on a three-month initial term and then monthly, since a month of testing rarely proves anything.
05How to start
From your numbers to the first change
Share the numbers
Analytics access, current conversion figures, traffic per key page, and who can change the site.
Research
Funnel analysis, session recordings, form data, support tickets and real user sessions show where people stop, and why.
Ranked backlog
Hypotheses ordered by expected impact and effort, with the required sample size calculated for each test.
Test or fix
Experiments run to completion, or direct changes where traffic is too low, reported the same way whether they won or lost.
Build the winners
Proven changes become proper code, so the testing tool does not remain a permanent layer on your site.
06For an accurate quote
What to include in your CRO request
Rough numbers are fine at this stage. We will verify them during the audit.
- The site URL and platform.
- Monthly sessions and conversions for the key pages.
- The step where you suspect people give up, and why you suspect it.
- Objections your support or sales team hears most often.
- Tools already installed for analytics, recordings or testing.
- Any redesign or platform change already planned.
07Questions
Questions buyers ask a CRO agency
How much does conversion rate optimisation cost?
It depends on the depth of research, whether experiments are feasible at your traffic level and how much development each change needs, so the quote follows a first look at your data. An audit with a ranked backlog is the natural first purchase, and it can stand alone.
How long does a CRO project take?
Plan on a research phase of a few weeks and then a continuous cycle, not a project with an end date. Each test runs until it reaches its calculated sample size, so low-volume pages take longer. A three-month initial term usually gives enough room for a first signal.
Can you guarantee a higher conversion rate?
No. Many well-reasoned tests lose, and conversion also depends on price, season and competition, which a website cannot control. What we commit to is the method: research first, sample sizes set in advance, and results reported as they are, including the losing ones.
We have low traffic. Is CRO still worth it?
It can be, but without A/B tests, because at low volume they give an unreliable verdict. We rely instead on user sessions, expert review and customer interviews. Treat the outcome as informed changes you then monitor, not as statistical proof that they worked.
Do we need a redesign to improve conversions?
Rarely as the first step, since research usually points at specific friction such as forms, delivery information or mobile errors. Fixing those is cheaper and shows what helped. A redesign makes sense when the platform or brand has to change anyway, or the evidence says so.
Will A/B testing hurt our SEO or page speed?
It should not if it is set up properly, though client-side testing tools do add some JavaScript. We end tests once they reach their sample size, follow search engine guidance on redirects and canonicals, watch Core Web Vitals during the test, and build winners into the site once a test ends.
Can your team implement the changes it proposes?
Yes, that is the point of having engineers and marketers together, so findings do not wait in someone else's queue. If your developers prefer to do the work, we write specifications and check the result in the same measurement set-up afterwards.
Request a quote
Describe what you need. The enquiry goes to our team, and we reply by email with the next step.